Home

Datalog Educational System V3.0 User`s Manual

1. Universidad Complutense de Madrid Datalog Educational System N S The predicate group_by admits a more compact representation than its SQL counterpart Let s consider the following Datalog session DES gt assert p 1 1 DES gt assert p 2 2 DES gt assert q X C group_by p X Y X C count C sum Y DES gt q X C Info Computing by stratum of p A B q 1 1 q 2 1 q 2 2 Info 3 tuples computed An analogous SQL session follows DES SQL gt create table p X int Y int DES SQL gt create view q X C as select X count Y as C from p group by X union select X sum Y as C from p group by X DES SQL gt select from q answer q X q C gt answer 1 1 answer 2 1 answer 2 2 Info 3 tuples computed 4 1 12 3 Aggregate Predicates An aggregate predicate returns its result in its last argument position as in sum p X X R which binds R to the cumulative sum of values for X provided by the input relation p These aggregate predicates simply allow another way of expressing aggregates in addition to the way explained just above Again with the same file the following queries are allowed DES gt count employee N D S S T Info Processing answer T count employee N D S S T answer 7 Info 1 tuple computed A group_by operation is simply specified by including the grouping variable s in the head of a clause as in the following view which
2. DES gt current_db Info The current database is Sdes DBMS S des DES gt assert p 1 DES gt assert p X r X Warning Undefined predicate s r 1 DES gt persistent p a int access Fernando Sdenz P rez 115 204 Universidad Complutense de Madrid Datalog Educational System DES gt p X p 1 p 2 p 3 Info 3 tuples computed DES gt r X Info 0 tuples computed DES gt use_db access DES gt current_db Info The current database is access DBMS access DES gt r X r 2 r 3 Info 2 tuples computed 5 2 9 2 Opening and Closing Connections Each time a persistent assertion is issued over a given connection this connection is opened although the current database is not changed to it In addition its is not closed although a drop_assertion command was issued A connection cannot be closed if any persistent predicate remains on it 5 2 9 3 Abolishing Predicates The command abolish not only abolishes rules in the deductive database but also those predicates that have been persistent in the external database dropping their table and view definitions 5 2 9 4 Null Values Processing of null values involving LDB and EDB is not still supported as they have different representations So outer joins are not supported up to now 5 2 9 5 External Database Processing Only the transferred rules of persisted predicates can be processed by the EDB In particular ne
3. Universidad Complutense de Madrid Datalog Educational System o Line numbers of the consulted programs are not reported for the source distribution of GNU Prolog since this system does not provide this information through read_term o GNU Prolog source distribution does not detect the ISO arithmetic error float_overflow int_overflow and int_underflow so that it is possible to get erroneous results when computations involve large numbers o GNU Prolog Windows application does not handle interactive command shells o GNU Prolog 1 4 0 does not seem to work on Windows XP SP3 error system_error error trying to execute pl2wam maybe not found consult 1 o Ciao source distribution does not support well SQL implementation as its FD constraint library is not complete enough for type checking o Ciao Prolog GNU Prolog and SWI Prolog distributions do not allow arithmetic expressions involving log 2 e ODBC issues o Neither Ciao Prolog nor GNU Prolog source distributions support ODBC connections 11 Release Notes This section lists release notes of the current DES version 11 1 Version 3 0 of DES released on May 10th 2012 e Enhancements o New commande close_db Name Close the given ODBC connection TAPI enabled drop_assertion Drop an assertion start_stopwatch Start stopwatch Precision depends on host Prolog system 1 second or milliseconds stop_stopwatch Stop stopwatch reset_stopwatch Reset s
4. answer 2 Info 1 tuple computed DES gt use_db S des DES gt select from q Error Unknown table or view q DES gt q X Warning Undeclared predicate s q 1 Fernando Sdenz P rez 112 204 Universidad Complutense de Madrid Datalog Educational System Info 0 tuples computed However a persisted predicate does have access to data and metadata in the EDB it was made persistent To show this and following the above system session let s assert the following rule DES gt assert p X q X Warning Undefined predicate s q 1 DES gt p X Info 0 tuples computed DES gt persistent p a int mysql DES gt p X p 2 Info 1 tuple computed Here the external database is assumed to hold a relation q 1 with a tuple q 2 in its meaning 5 2 8 Applications Persisting predicates opens a brand new scenario for several reasons First predicates are no longer limited by available memory instead persisted predicates are using as much secondary storage as needed and provided by the underlying external database Predicate size limit is therefore moved to the external database Second processing is directed to the external database for rules that can be projected and to the deductive engine for rules that can not This way one can take advantage of the external database performance and scalability Third queries which are not possible in an external database can be solved by the
5. safe X X U V opp X X1 state X1 X1 U V Farmer takes Goat state X Y X V 18 Adapted from Diet87 Fernando Sdenz P rez 177 204 Universidad Complutense de Madrid Datalog Educational System safe X Y X V opp X X1 state X1 Y X1 V Farmer takes Cabbage state X Y U X safe X Y U X opp X X1 state X1 Y U X1 Farmer goes by himself state X Y U V safe X Y U V opp X X1 state X1 Y U V Opposite shores n s opp n s opp s n Farmer is with Goat safe X Y X V Farmer is not with Goat safe X X X1 X opp X X1 If we submit the query state s s s s we get the expected result state s s s s Info 1 tuple computed That is the system has proved that there is a serial of transfers between shores which finally end with the asked configuration this problem is not modeled to show this serial If we ask for the extension table contents regarding the relation state 4 with the command list_et state 4 we get for the answers state n n n n state n n n s state n n s n state n s n n state n s n s state s n s n state s n s s state s s n s state s s s n state s s s s Info 10 tuples in the answer set This is the complete set of valid states which includes all of the valid paths from state n n n n to state s s s s However the order of states to reach the latter is not given but we can find it by obse
6. 8 Universidad Complutense de Madrid Datalog Educational System Appendix A License A 1 Software License DES licensing comes from the ideas of the Free Software Foundation Since version 3 0 it is distributed under version 3 of the GNU Lesser General Public License LGPL which supplements version 3 of the GNU General Public License DES is free software you can redistribute it and or modify it under the terms of the GNU General Public License as published by the Free Software Foundation either version 3 of the License or at your option any later version DES is distributed in the hope that it will be useful but WITHOUT ANY WARRANTY without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE See the GNU General Public License for more details You should have received a copy of the GNU General Public License along with this program If not see http www gnu org licenses DES versions prior to 3 0 were distributed under GNU General Public License GPL A 2 Documentation License GNU Free Documentation License Version 1 3 3 November 2008 Copyright 2000 2001 2002 2007 2008 Free Software Foundation Inc lt http fsf org gt Everyone is permitted to copy and distribute verbatim copies of this license document but changing it is not allowed 0 PREAMBLE The purpose of this License is to make a manual textbook or other functional and useful document free in the sense
7. Checking that a view produces the same result as its intended interpretation is a daunting task when large databases and both dependent and correlated queries are considered Test case generation provides tuples that can be matched to the intended interpretation of a view and therefore be used to catch possible design errors in the view A test case for a view in the context of a database is a set of tuples for the different tables involved in the computation of the view Executing a view for a positive Fernando Sdenz P rez 134 204 Universidad Complutense de Madrid Datalog Educational System test case DICH should return at least one tuple This tuple can be used by the user to catch errors in the view if any This way if the user detects that this tuple should not be part of the answer it is definitely a witness of the error in the design of the view On the contrary the execution of the view for a negative test case NTC should return at least one tuple which should not be in the result set of the query Again if no such a tuple can be found this tuple is a witness of the error in the design A PTC in a basic query means that at least one tuple in the query domain satisfies the where condition In the case of aggregate queries a PTC will require finding a valid aggregate verifying the having condition which in turn implies that all its rows verify the where condition In the case of basic query a NTC will contain a
8. DES gt 2 1 2 4 Mac OS X From the same URL above you can download a Mac OS X executable distribution in a single archive file containing the following readmeDES lt version gt A quick installation guide and file release contents Datalog Educational System des Console executable file doc manualDES lt version gt pdf This manual examples dl Example files which will be discussed in Section 6 license license A verbatim copy of the GNU Public License for this distribution The following screenshot has been taken in Mac OS X Snow Leopard e00 Terminal swipl 66x21 svps Mac des3 0 svp des REE OE EE a oo oa a eo a oe oe a a a a oe a ne a oO oo ok oe a a Oo DES Datalog Educational System v 3 0 Type help for help about commands Fernando Saenz Perez c 2004 2012 GPD DISIA UCM Please send comments questions etc to fernan sip ucm es Web site http des sourceforge net SS SS SS SS SS RS Ke EES This program comes with ABSOLUTELY NO WARRANTY is free software and you are welcome to redistribute it under certain conditions Type license for details oo oo a a a a aa ao ao aa hho a aaa ob DES gt J Fernando S enz P rez 13 204 Universidad Complutense de Madrid Datalog Educational System 2 2 Installing and Executing DES Unpack the distribution archive file into the directory you want to install DES which will be referred to as the distribution
9. Fernando Sd enz P rez 53 204 Universidad Complutense de Madrid Datalog Educational System Referenced columns have to match the types of foreign key columns otherwise an error is raised DES gt fk r c q b Error Type mismatch r c string varchar lt gt q b number integer A relation already defined with facts or rules is checked for consistency when trying to assert a new foreign key constraint DES gt type p a int DES gt type q a int DES gt assert p 1 DES gt pk q a DES gt fk p a q a Error Foreign key violation p a gt q a Offending values in database fk 1 Info Constraint has not been asserted 4 1 14 6 Functional Dependency A functional dependency constraint specifies that given a set of attributes Ai of a relation R they functionally determine another set Az i e each tuple of values of Ai in R is associated with precisely one tuple of values Az in the same tuple of R DES gt fd p a Tel Error Relation p has not been typed yet DES gt type p a int b int DES gt fd p a Tel Error Unknown column c DES gt fd p a b DES gt dbschema p Info Table p a number integer b number integer FD a gt b By asserting the fact p 1 2 it must hold that any other tuple with 1 in its first attribute must have the value 2 in its second attribute DES gt assert p 1 2 DES gt assert p 1 3 Error
10. INSERT INTO TableName Att Att VALUES Cte Cte INSERT INTO TableName Att Att DQLstmt DELETE FROM TableName DELETE FROM TableName WHERE Condition UPDATE TableName SET Att1 Exprl1 Attn Exprn WHERE Condition Fernando Sdenz P rez 77 204 Cte is a constant LESESEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEES DQL Data Query Language statements LESEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEES DOLstmt DQLstmt UBSQL UBSQOL SELECTstmt Deen UNION ALL DQLstmt pitas EXCEPT DOLstmt ee MINUS DQLstmt en INTERSECT DQLstmt ae LocalViewDefinition LocalViewDefinition DQLstmt ASSUME LocalAssumption LocalAssumption DQLstmt LocalViewDefinition RECURSIVE CompleteSchema AS DQLstmt LocalAssumption DOLstmt IN CompleteSchema SELECTstmt SELECT TOP Integer ALL DISTINCT SelectExpressionList FROM Rels WHERE WhereCondition GROUP BY Atts HAVING HavingCondition ORDER BY OrderDescription FETCH FIRST Integer ROWS ONLY Atts rs Att Att OrderDescription Att ASC DESC Att ASC DESC SelectExpressionList SelectExpression SelectExpression SelectExpression Universidad Complutense de Madrid Datalog Educational System Fernando Sdenz P rez 78 204 Universidad Complutense de Madrid Datalog Educational System UnrenamedSelectExpression RenamedExpression UnrenamedSelectExpression A
11. On the one hand the number of positive facts which can be inferred are finite because there is a finite number of ground facts which can be used in a given proof and proofs have finite depth provided that tabling prevents recomputations of older nodes in the proof tree On the other hand the number of negative facts which can be inferred is also finite because they are proved using negation as failure Failures are always finite because they are proved trying to get a success Finally there are facts that cannot be proved to be true or false because of recursion These cases are detected by the tabling mechanism which prevent infinite recursion such as inp p It is also possible that both a positive and a negative fact have been inferred for a given call Then an undefined fact replaces the contradictory information The implementation simply removes the contradictory facts and informs about the undefinedness As already indicated see Section 6 9 the algorithm for determining undefinedness is incomplete 5 16 3 Dependency Graphs and Stratification Negation Outer Joins and Aggregates Each time a program is consulted or modified Oe via submitting a temporary view or changing the database a predicate dependency graph is built ZCF 97 This graph shows the dependencies through positive and negative atoms among predicates in the program Also a negative dependency is added for each outer join goal and aggregate goal This dep
12. and subsumed by the current one which is a brand new implementation There possible erroneous objects correspond to views and the debugger looks for erroneous views asking the user whether the result of a given view is as expected When the user starts the debugger for a view with the command debug_sql View the debugger builds internally its computation tree and starts the debugging session The root of the tree is the view under debugging its nodes can be either views or tables and children of a view are all of the views and tables occurring in that view table nodes do not have children This tree is traversed and the validity whether the view outcome matches its intended meaning of each node is asked to the user If a given node is checked as valid its subtree is assumed to be valid and it is no longer traversed Otherwise the node itself or one of its descendants is assumed to be nonvalid In this case the subtree is traversed to find the erroneous node Considering the file pets1 sql in the directory examples SQLDebugger the problem is explained in the same file we find that the view Guest returns an unexpected answer DES gt process examples SQLDebugger pets1 sql DES gt select from Guest answer Guest id number integer Guest name string varchar 50 gt answer 1 Mark Costas answer 2 Helen Kaye answer 3 Robin Scott Info 3 tuples computed In fact only Robin Scott is exp
13. d 0 Input Continue y n yl Info Tracing predicate d Info No more predicates to trace 5 7 2 Tracing SQL Views Tracing SQL views is similar to tracing Datalog queries but instead of posing a goal involving in general variables and constants to trace only the name of a view should be given For example let s consider the file family sql which contains view definitions for ancestor and parent where tables father and mother are involved in the latter view Note that this view is recursive since it depends on itself create view parent parent child as select from father union select from mother create or replace view ancestor ancestor descendant as select parent child from parent union select parent descendant from parent ancestor where parent child ancestor ancestor Then tracing the view ancestor is as follows DES SQL gt trace_sql ancestor Info Tracing view ancestor ancestor amy carolIII Fernando Sdenz P rez 126 204 Universidad Complutense de Madrid Datalog Educational System ancestor tony carolIIT Info 16 tuples in the answer table Info Remaining views parent 2 father 2 mother 2 Input Continue y n yl Info Tracing view parent parent amy fred parent tony carolIT Info 8 tuples in the answer table Info Remaining views father 2 mother 2 Input Continue y n yl Info Tracing view father father f
14. des p1 of the system making it incompatible with other platforms This is why the core file for Ciao has some preliminary directives not found in the core file shared by other platforms Future Ciao versions may change this particular behaviour GNU Prolog as well needs a prelude for avoiding the initialization call to ensure_loaded 1 since it does not support this ISO predicate See also Section 10 for consult unsupported features of some source distributions 6 Examples The DES distribution contains the directory examples which shows several features of the system Unless explicitly noted all queries have been solved after the commands verbose off and pretty_print off have been executed Fernando Sdenz P rez 166 204 Universidad Complutense de Madrid Datalog Educational System 6 1 Relational Operations files relop d1 sql rail The program relop dl is intended to show how to mimic with Datalog rules the basic relational operations that can be found in the file relop sql It contains three relations a b and c which are used as arguments of relational operations In order to have loaded this program and be able to submit queries you can consult it with c relop In the remarks below relational operator symbols are represented with ASCII characters as x to denote the left outer join X and x to simply denote the Cartesian product Extended Relational Algebra Operations pi X c X Y Projecti
15. the existence operator EXISTS and the inclusion operator IN See the grammar description in Section 4 2 8 for details Subqueries are allowed with no limitations Relations is a list of comma separated relation definitions A relation can be either a table name or a view name or a subquery or a join relation They can be renamed via aliases If no FROM clause is provided the built in DUAL relation is used as a data source cf Section 4 2 6 1 2 Examples Given the tables CREATE TABLE s a int b int CREATE TABLE t a int b int CREATE TABLE v a int b int We can submit the following queries SELECT distinct a FROM t SELECT t s b FROM t s v WHERE t a s a AND v b t b SELECT t a s b t ats b FROM t s WHERE t a s a SELECT FROM SELECT from t as rl SELECT from s as r2 WHERE r1 a r2 b Fernando Sdenz P rez 67 204 Universidad Complutense de Madrid Datalog Educational System SELECT FROM s WHERE s a NOT IN SELECT a FROM t SELECT FROM s WHERE EXISTS SELECT a FROM t WHERE t a s a SELECT FROM s WHERE s a gt SELECT a FROM t SELECT 1 alta2 atl AS al a 2 AS a2 FROM t SELECT 1 Notes e SQL arithmetic expressions follow the same syntax as Datalog e A SQL arithmetic expression can be renamed and used in other expressions e Circular definitions will yield exceptions at run time as in a a3 AS a3 A join relation is either of the form Relation NATURA
16. 1 1 1 0 1 2E34 1 2E 34 and 1 2E 34 Note that 1 1 1 1 E23 and 1E23 are not valid numbers A plus sign is not part of a positive number however a minus sign can be used as a prefix unary operator in arithmetical expressions cf Section 4 5 4 1 and also following the symbol E in scientific notation as already seen e Constants A constant can be o A number integer or float o Any sequence of alphanumeric characters including the underscore _ starting with a lowercase letter o Any sequence of characters delimited by single quotes Examples of alphanumeric constants are foo foo_foo foo foo 2 3 and X e Variables Variables are written with alphanumeric characters and alternatively start with either an uppercase or with an underscore _ Anonymous variables are also allowed which are denoted with a single underscore Each occurrence of an anonymous variable is considered different from any other anonymous variable For instance in the rulea b _ c _ both goals do not share variables Any variable starting with an underscore either anonymous or not is removed from a computed query cf Section 4 1 7 Examples of variables are X _X wear and _ e Unknowns Unknowns are represented as null values and are written alternatively as both null and NULL ID where ID is a unique identifier The first form is used for normal users whilst the second one is intended for development uses cf
17. 3 rules listed DES gt 1 X Y 1 1 1 1 2 SNULL 72 1 SNULL 70 SNULL 74 Info 3 tuples computed Observe above ID 70 There the data source rule providing such an entry in the answer is the first rule of p As SQL statements and RA expressions are compiled to Datalog programs the command show_compilations on enables the display of compilations each time a SQL statement is submitted as the following example illustrates DES gt show_compilations on DES gt create table t a int b int DES gt create table s a int b int DES gt select from t where a gt 1l union select from s where b lt 2 Info SQL statement compiled to answer A B distinct answer_2_1 A B answer_2_1 A B t A B A gt 1 answer_2_1 A B s A B B lt 2 answer t a t b gt Fernando S enz P rez 124 204 Universidad Complutense de Madrid Datalog Educational System Info 0 tuples computed 5 7 Datalog and SQL Tracers In contrast to imperative programming languages deductive and relational database query languages feature solving procedures which are far from the query languages itself Whilst one can trace an imperative program by following each statement as it is executed along with the program state this is not feasible in declarative high abstraction languages as Datalog and SQL However this does not apply to Prolog also acknowledged as a declarative language beca
18. Also allowed the alternative syntax CANDIDATE KEY REFERENCES TableName Column Referential integrity constraint for only one column Check constraints are not supported in this syntax up to now However they can be imposed via Datalog user defined constraints as explained in Section 4 1 14 7 Also there is provision for several table constraints e PRIMARY KEY Column Column Primary key constraint for one or more columns e UNIQUE Column Column Uniqueness constraint for one or more columns Also allowed the non standard alternative syntax CANDIDATE KEY Column Column Fernando Sdenz P rez 61 204 Universidad Complutense de Madrid Datalog Educational System e FOREIGN KEY Column Column REFERENCES TableName Column Column Referential integrity constraint for one or more columns Allowed types include e CHAR Fixed length string of 1 e CHAR n Fixed length string of n characters e VARCHAR n Variable length string of up to n characters e VARCHAR or STRING Variable length string of up to the maximum length of the underlying Prolog atom e INTEGER or INT Integer number e REAL Real number Examples CREATE TABLE t a INT PRIMARY KEY b STRING CREATE OR REPLACE TABLE s a INT b INT REFERENCES t a PRIMARY KEY a b Note in this last example that if the column name in the referential integrity constraint is missing the referred column of table t is assumed t
19. Example Input tapi relation_exists v Output true Command tapi ddl_query Answer Regular Remarks This DDL statement returns success upon a successful processing Example Input tapi create table t a int Output Ssuccess Command tapi dependent_relations pattern Fernando Sdenz P rez 155 204 Universidad Complutense de Madrid Datalog Educational System Where pattern can be either relation_name or relation_name arity where relation_name stands for a relation name and arity for its arity Answer relation_name relation_name Seot Where relation_name stands for relation names Remarks Display the names of relations that directly depend on the given relation Relations are returned alphabetically sorted Example Input considering that views z1 y z2 reference table t tapi dependent_relations t Output z1 ER Seot e Command tapi list_table_ schemas Answer table_name column_name type column_name type table_name column_name type column_name type table_name column_name type column_name type Seot Where table_name stands for table names column_name is a column name type is the column type and eot is the end of the transmission Remarks Return table schemas Tables are returned alphabetically sorted Example Input tapi list_table_schemas Output t a number integer Seot e Command tapi list_view_schemas Answ
20. Example a difference b Cartesian Product Ri x Ro Concrete syntax Relation1 product Relation2 Example a product b GEES Concrete syntax rename Schema Relation Example project v a rename v a select true a Assignment R A1 An lt Ro Create a new relation Ri with argument names Al An as a copy of Ro It allows to define new views Concrete syntax Relationl Relation2 Example viel select true a 4 3 1 2 Additional operators These operators can be expressed in terms of basic operators and include Set Intersection Ri A R2 Concrete syntax Relationl1 intersect Relation2 Example a intersect b x Theta join R R2 Equivalent to oR x R Concrete syntax Fernando Sdenz P rez 83 204 Universidad Complutense de Madrid Datalog Educational System Relation1 zjoin Condition Relation2 Example a zjoin a a lt b b b bd Natural inner join Ri R2 Return tuples of Ri joined with R2 such that common attributes are pair wise equal and occur only once in output relation Concrete syntax Relation1 njoin Relation2 Example a njoin cC 4 3 1 3 Extended operators These operators can not be expressed in terms of former operators and include Extended projection 7 1 en R Return tuples of R with columns F1 E where each E is an expression built from constants and attributes of R Concrete syntax project El En Relat
21. Functional dependency violation p a gt p b in table p a b when trying to insert p 1 3 Witness tuple p 1 2 Several functional dependency constraints can be imposed on a given relation They can be deleted either with the command drop_ic or when a SQL DROP TABLE or DROP DATABASE statements are issued Trivial functional dependencies are rejected DES gt fd p a a Warning Trivial functional dependency Not asserted Fernando Sdenz P rez 54 204 Universidad Complutense de Madrid Datalog Educational System A relation already defined with facts or rules is checked for consistency when trying to assert a new functional dependency constraint DES gt type p a int b int c int DES gt assert p 1 1 1 DES gt assert p 1 2 3 DES fd p a Tel Error Functional dependency violation p a gt p c Offending values in database fd 1 1 1 fd 1 2 3 Info Constraint has not been asserted 4 1 14 7 User defined Integrity Constraints Users can also define their own integrity constraints A user defined integrity constraint is represented with a rule without head The rule body is an assertion that specifies inconsistent data i e should this body can be proved an inconsistency is detected and reported to the user Declaring such integrity constraints implies to change your mind w r t usual consistency constraints as domain constraints in SQL For instance to specify that a column c
22. Note that if you have previously created some database objects tables views in DES without an ODBC connection they are still available and can be queried system asks for dropping the current DES database before starting to use a new database 5 1 2 Using a Connection Assuming that the connection links to an empty database let s start creating some database objects Note that depending on the installed MySQL ODBC driver version annoying messages can be displayed DES SQL gt create table t a varchar 20 primary key DES SQL gt create table s a varchar 20 primary key DES SQL gt create view v a b as select fromt s DES SQL gt insert into t values 1 Info 1 tuple inserted DES SQL gt insert into s select from t Info 1 tuple inserted DES SQL gt insert into s values 2 Info 1 tuple inserted Next one can ask for the database schema metadata with the command DES SQL gt dbschema Info Table s s a varchar t a varchar Info View s v a varchar b varchar All of these tables and views can be accessed from DES as if they were local DES SQL gt select from s answer a varchar gt answer 1 answer 2 Info 2 tuples computed DES SQL gt select from t answer a varchar gt answer 1 6 Further improvements of the system will include to handle multiple database connections removing the requirement of dropping the DES database Fernando
23. X Y C group_by v X Y X Y C count answer 1 1 2 answer 1 2 2 Info 2 tuples computed Fernando Sdenz P rez 97 204 Universidad Complutense de Madrid Datalog Educational System Note that even when you can access SQL objects from Datalog the contrary is not allowed because there is nor Datalog metadata information for the external SQL engine neither access to Datalog data The data bridge is only opened from DES to the external DBMS not the other way round This is in contrast to the SQL database internally provided by DES which allows a bidirectional communication since type information is supported for Datalog predicates 5 1 3 Opening Several Connections From release 3 0 on several OCBC connections can be opened simultaneously Each time a new connection is opened it becomes the new current connection and all query processing is related to it by default For instance to inspect a rather limited set of metadata one can submit the following command DES gt open_db mysql DES gt dbschema Info Database mysql Info Table s s a varchar 20 t a integer 4 w a varchar 20 Info View s v a varbinary 20 Info No integrity constraints To list all the opened connections use the command DES gt show_dbs Sdes access csv db2 excel mysql oracle postgresql sqlserver where you can see the list of opened connections starting with des
24. editors CIMNE pp 41 50 Barcelona Spain September 2006 R Caballero Y Garcia Ruiz and F Sdenz P rez A New Proposal for Debugging Datalog Programs 16th International Workshop on Functional and Constraint Logic Programming 2007 R Caballero Y Garcia Ruiz and F Sdenz P rez A Theoretical Framework for the Declarative Debugging of Datalog Programs In International Workshop on Semantics in Data and Knowledge Bases SDKB 2008 LNCS 4925 pp 143 159 Springer 2008 R Caballero Y Garcia Ruiz and F Sdenz P rez Applying Constraint Logic Programming to SQL Test Case Generation In 10th Fernando Sdenz P rez 201 204 Universidad Complutense de Madrid Datalog Educational System CGS11b CGS12a Chan78 Diaz Diet87 Diet01 DMP93 Drax92 FD92 FHH04 FP96 GR68 GTZ05 GUW02 HA92 International Symposium on Functional and Logic Programming FLOPS 2010 2010 R Caballero Y Garc a Ruiz and F S enz P rez Algorithmic Debugging of SQL Views Eigth Ershov Informatics Conference PSI 11 Novosibirsk Akademgorodok Russia June 2011 R Caballero Y Garcia Ruiz and F Sd enz P rez Declarative Debugging of Wrong and Missing Answers for SQL Views In 11th International Symposium on Functional and Logic Programming FLOPS 2012 Springer Lecture Notes in Computer Science Kobe Japan May 2012 C L Chang Deduc
25. facts The current implementation uses an incomplete algorithm for finding such undefined facts We can see this incompleteness by adding the following rule shaved M shaves barber M The query shaved M returns Warning Unable to ensure correctness for this query shaved mayor Info 1 tuple computed That is the system is unable to prove that shaved barber is undefined If you look at the predicate dependency graph and the stratification of the program DES gt pdg Nodes man 1 shaved 1 shaves 2 Arcs shaves 2 shaves 2 shaves 2 man 1 shaved 1 shaves 2 DES gt strata non stratifiable Fernando Sdenz P rez 180 204 Universidad Complutense de Madrid Datalog Educational System you get the predicate dependency graph shown in Figure 4 and you are informed that the program is non stratifiable This figure shows a negation in a cycle so that the program is not stratifiable The system warned of this situation when the program was loaded shaves 7 X man shaved Figure 4 Predicate Dependency Graph for russell dl However even when a program is non stratifiable there may exist a query with an associated predicate dependency subgraph so that negation does not occur in any cycle For instance this occurs with the query man X in this program DES gt man X Info Stratifiable subprogram found for the given query man barber man mayor Info 2 tuples computed
26. i e ignoring nulls Duplicate free counterparts are also provided sum_distinct count_distinct avg_distinct and times_distinct Note that for minimum and maximum no counterparts are provided since they would compute the same results 4 1 12 1 Aggregate Functions An aggregate function can occur in expressions and returns a value as in R 1 sum X where sum is expected to compute the cumulative sum of possible values for X and X has to be bound in the context of a group_by predicate cf next section wherein the expression also occur 4 1 12 2 Group_by Predicate A group_by predicate encloses a query for which a given list of variables builds answer sets groups for all possible values of these variables Let s consider the following excerpt from the file aggregates dl employee Name Department Salary employee anderson accounting 1200 Fernando Sdenz P rez 42 204 Universidad Complutense de Madrid Datalog Educational System employee andrews accounting 1200 employee arlingon accounting 1000 employee nolan null null employee norton null null employee randall resources 800 employee sanders sales null employee silver sales 1000 employee smith sales 1000 employee steel sales 1020 employee sullivan sales null We can count the number of employees for each department with the following query DES gt group_by employee N D S D R count Info Pro
27. p X X 1 X 2 Info 1 rule listed DES gt persistent p a int mysql DES gt listing p X X 1 XS 2 Info 1 rule listed DES gt quit Then we open a new system session and type DES gt persistent p a int mysql Info Recovering existing data from external database DES gt listing p A A 2 p A A 1 Info 2 rules listed As can be seen two rules are the result of the compilation of the originally asserted single rule with a disjunctive body Also original variable names only X in Fernando Sdenz P rez 108 204 Universidad Complutense de Madrid Datalog Educational System tnis case are missing However a next release of DES might deal with this allowing to restore the very same rules as the original ones 5 2 5 Schema of Persistent Predicates You can request the current database schema with DES gt dbschema Info Database Sdes Info No tables Info View s p a number integer Defining SQL statement CREATE VIEW p a AS SELECT ALL FROM p_des_table Datalog equivalent rules Info No integrity constraints where the persisted predicate is listed in the database schema of the default database des and therefore it can be combined in a query with any predicate visible in this database Note that predicate p has been declared as a view depending on a table with the same name as the predicate and view but ending with _des_table
28. p0 1 SNULL 0 Info 3 tuples in the answer table Calls p0 A B Info 1 tuple in the call table Extension table contains the non ground entry not p1 1 A which is not safe 5 4 Source to Source Transformations Currently two source to source transformations are possible under demand First as explained in the previous section when safety transformations are enabled via the command safe on rule bodies are reordered to try to produce a safe rule Second when simplification is enabled via the command simplification on rule bodies containing equalities true and not BooleanValue are simplified In addition there is also place for several automatic transformations cf Section 5 6 to know how to display such transformations e A clause containing a disjunctive body is transformed into a sets of clauses with conjunctive bodies e A clause containing an outer join predicate is transformed into an executable form e A clause containing an aggregate predicate is transformed into an executable form including grouping criterion e A clause containing the goal not is_null Term is transformed into a clause with this goal replaced by is_not_null Term Fernando Sdenz P rez 121 204 Universidad Complutense de Madrid Datalog Educational System 5 5 Multi line Mode By default DES command prompt reads single line inputs and therefore ending termination character is optional
29. select a from a inner join b on a a b b Left Join select from a left join b on a a b b Right Join select from a right join b on a a b b Full Join select from a full join b on a a b b Union select from a union select from b Difference select from a except select from b If we have created the relations in Datalog we cannot access them from SQL unless they had been either defined as tables or views or declared with types For example following the first alternative and after consulting the file relop d1 we can submit create table a a varchar And then accessing with a SQL statement the tuples that were asserted in Datalog DES SQL gt select from a answer a a gt answer al answer a2 answer a3 Info 3 tuples computed Otherwise an error is submitted Error Unknown table or view a Following the second alternative and after consulting the file relop dl we can declare types for a DES SQL gt datalog type a a varchar DES SQL gt select from a answer a a gt answer al answer a2 answer a3 Info 3 tuples computed Fernando S enz P rez 169 204 Universidad Complutense de Madrid Datalog Educational System 6 2 Paths in a Graph files paths d1 sql ra This program introduces the use of recursion in DES by defining the graph in Figure 1 and the set of tuples lt origin destination gt such that there is a path f
30. the title equally prominent and visible You may add other material on the covers in addition Copying with changes limited to the covers as long as they preserve the title of the Document and satisfy these conditions can be treated as verbatim copying in other respects If the required texts for either cover are too voluminous to fit legibly you should put the first ones listed as many as fit reasonably on the actual cover and continue the rest onto adjacent pages If you publish or distribute Opaque copies of the Document numbering more than 100 you must either include a machine readable Transparent copy along with each Opaque Fernando Sdenz P rez 195 204 Universidad Complutense de Madrid Datalog Educational System copy or state in or with each Opaque copy a computer network location from which the general network using public has access to download using public standard network protocols a complete Transparent copy of the Document free of added material If you use the latter option you must take reasonably prudent steps when you begin distribution of Opaque copies in quantity to ensure that this Transparent copy will remain thus accessible at the stated location until at least one year after the last time you distribute an Opaque copy directly or through your agents or retailers of that edition to the public It is requested but not required that you contact the authors of the Document well before redistributi
31. 151 5 14 2 TAPI enabl d Commands vc sissioni niina 151 5 14 3 TA Pl enabled Queries nanmin ian Wate tae ees 160 5 15 ISO Escape Character Syntax ee EE 162 5 16 Notes about the Implementation of DES iciccctisincectorcntenteranctorsrencavrdesutsntasiitaserad 163 DE 164 5 16 2 Pixpoint Compt tat OM vesie anaa eee EE AEE EEEE 165 5 16 3 Dependency Graphs and Stratification Negation Outer Joins and PR OOD AL OS aree e E et ve E et see RS E EE 165 5 16 4 Porting to Unsupported Systems een 166 5 16 5 Differences among PIA ORM aerer Eder 166 fe EELER 166 6 1 Relational Operations files reLop d1 SQ1 a cecsssccsssseeetessseeenees 167 6 2 Paths in a Graph files paths dl sql a sistssccsssveisssecveccsserarentavavesesicuvies 170 6 3 Shortest Paths file spaths d1 Sql xa EE 171 6 4 Family Tree files family qd1 sql Ea EK eet stewbia diustinas 173 6 5 Basic Recursion Problem file recursion dl ss ssesssssssrreeririrererrrereresese 175 6 6 Transitive Closure files tranclosure d1 sql rail 175 6 7 Mutual Recursion files mutrecursion d1l sql rail 176 6 8 Farmer Wolf Goat Cabbage Puzzle file puzzle dl 177 6 9 Paradoxes files russell d1 Sq rail 179 610 Rarity tile EE Ee e 182 6 11 Grammar fileo rari an os aeclactnaialaseeaeceevnatorrorted nee eee ovens 183 6 12 Fibonacet file fib dl Sq Pat neer ar Manin E i 183 6 13 Hanoi Towers file Eege EE aerer E EE ee 184 6 14 Other Exa MpleS doree eE a esate
32. 2 awards Mica Input Is this the expected answer for view awards y n m mT w wN a h n wl Info Debugging view intensiveStudents 1 intensiveStudents Juan Input Is this the expected answer for view intensiveStudents y n m mT w wN a h y Info Debugging view candidates Input Should candidates include a tuple of the form Ana y n a y n Info Debugging view basicLevelStudents Input Should basicLevelStudents include a tuple of the form Ana y n a y n Info Debugging view salsaStudents Input Should salsaStudents include a tuple of the form Ana 1 teach1 y n a y Info Debugging view salsaStudents Input Should salsaStudents include a tuple of the form Ana 2 teach2 y n a y Info Debugging view salsaStudents Input Should salsaStudents include a tuple of the form Ana 3 teach1 y n a y Info Buggy view found basicLevelStudents 5 9 2 3 Displaying Extended Information Enabling verbose output allows to extend the display with further information as e g view definitions when they are asked for its validity As well enabling development output allows to check how the logic program that represents the computation tree is built c f CGS12a For that use the following commands resp DES gt verbose on Info Verbose output is on DES gt development on Info Development listings are on 5 10 SOL Test Case Generator
33. Base Theory Volume 1 Plenum Press 1981 J Lloyd Foundations of Logic Programming Springer Verlag 1987 J Minker Perspectives in Deductive Databases Technical Report CS TR 1799 University of Maryland at College Park March 1987 J Minker and J M Nicolas On Recursive Axioms in Deductive Databases Information Systems 16 4 670 702 1991 J Matuszynski and A Sza as Living with Inconsistency and Taming Nonmonotonicity To appear in Datalog 2 0 G Gottlob G Grasso O de Moor and A Sellers eds LNCS 6702 334 398 Springer Verlag 2011 G Phipps M A Derr and K A Ross Glue NAIL A Deductive Database System In Proc of the ACM SIGMOD Conference on Management of Data pp 308 317 1991 J A Robinson A Machine Oriented Logic Based on the Resolution Principle Journal of the ACM 12 23 41 1965 R Ronen and O Shmueli Evaluating very large Datalog queries on social networks In EDBT 09 Proceedings of the 12th International Conference on Extending Database Technology pages 577 587 New York NY USA 2009 ACM R Ramakrishnan D Srivastava S Sudarshan and P Seshadri The Coral deductive system VLDB Journal 3 2 161 210 1994 P Rao Konstantinos F Sagonas Terrance Swift David Scott Warren and Juliana Freire XSB A System for Efficiently Computing WFS Logic Programming and Non monotonic Reasoning 1997 R Ramakrishnan and J D Ullman A Surv
34. Info 3 tuples computed DES SQL gt select from t answer a varchar gt answer 1 answer 2 Info 2 tuples computed This reveals that although on the DES side Datalog data are known it is not on the RDBMS side This is in contrast to the DES management of data if no ODBC connection is opened the DES engine is aware of any changes to data both from Datalog and SQL sides Concluding those updates that are external to DES might not be noticed by the DES engine And also an ODBC connection should be seen as a source of external data that should not be mixed with Datalog data However you can safely use the more powerful Datalog language to query external data and to be sure the current data is retrieved clear the cache with clear_et 5 1 9 2 ODBC Metadata When computing the predicate dependency graph and stratification metadata from the external DBMS is retrieved which can be a costly operation if the number of tables and views is large This is the default case when opening connections to DBMSs as SQL Server or Oracle where many views are defined for an empty database Also ODBC connections to Oracle seem to be slow Listing the database schema can suffer this situation as well by issuing the command dbschena Instead it is better to focus on the required object to display as either dbschema relname or dbschema connection relname 5 1 9 3 ODBC Limitations As predicate dependency graphs are
35. OS X RDBMSs include enterprise RDBMS as Oracle MySQL DB2 and desktop RDBMS as MS Access and FileMaker ODBC drivers are usually bundled with OS platforms as Windows OSs ODBC implementation Linux OS distributions as Ubuntu Red Hat and Mandriva UnixODBC implementation and Mac OSs 10x ODBC implementation However additional drivers for specific databases are needed to be installed Since each RDBMS provides an ODBC driver and each OS an ODBC implementation details on how to configure such connections are out of the scope of this manual However to configure such a connection typically the ODBC driver is looked for and installed in the OS Then following the manufacturer recommendations it is configured You can find many web pages with advice on this Here we assume that there are ODBC connections already available 5 1 1 Opening an ODBC Connection To access a RDB in DES first open the connection with the following command where test is the name of a previously created ODBC connection to a database DES SQL gt open_db test You can also provide a username and password if needed as in DES SQL gt open_db test user smith password my_pwd Fernando Sdenz P rez 94 204 Universidad Complutense de Madrid Datalog Educational System If ODBC connector returns an error then you have to enclose these between apostrophes as in DES SQL gt open_db test user smith password my_pwd
36. Resolution with Tabulation Proceedings of ICLP 86 Lecture Notes on Computer Science 225 Springer Verlag 1986 J D Ullman Database and Knowledge Base Systems Vols I Classical Database Systems and II The New Technologies Computer Science Press 1995 J Vaghani K Ramamohanarao D B Kemp Z Somogyi and P J Stuckey Design Overview of the Aditi Deductive Database System In Proc of the 7th Intl Conf on Data Engineering pp 240 247 1991 J Wielemaker http www SWI Prolog org J Whaley and M Lam Cloning based context sensitive pointer alias analyses using binary decision diagrams In Prog Lang Design and Impl 2004 C Zaniolo S Ceri C Faloutsos T T Snodgrass V S Subrahmanian and R Zicari Advanced Database Systems Morgan Kauffmann Publishers 1997 U Zukowski and B Freitag The Deductive Database System LOLA In J Dix and U Furbach and A Nerode Eds Logic Programming and Nonmonotonic Reasoning LNAI 1265 pp 375 386 Springer 1997 Fernando Sdenz P rez 204 204
37. Sdenz P rez 95 204 Universidad Complutense de Madrid Datalog Educational System Info 1 tuple computed DES SQL gt select from v answer a varchar b varchar gt answer 1 1 answer 1 2 Info 2 tuples computed DES SQL gt insert into t values 1 Exception error odbc 23000 1062 MySQL ODBC 3 51 Driver mysqld 5 0 41 community nt Duplicate entry 1 for key 1 _G3 In this example as table t has its single column defined as its primary key trying to insert a duplicate entry results in an exception from the ODBC driver Integrity constraints are handled by the RDBMS connected instead of DES notice that the exception message is different from the one generated by DES Moreover you can submit SQL statements that are not supported by DES but otherwise by the connected RDBMS as DES SQL gt alter table t drop primary key Then you can insert again and see the result including duplicates DES SQL gt insert into t values 1 Info 1 tuple inserted DES SQL gt select from v answer a varchar b varchar gt answer 1 1 answer 1 1 answer 1 2 answer 1 2 Info 4 tuples computed Also duplicate removing is also possible by the external RDBMS DES SQL gt select distinct from v answer a varchar b varchar gt answer 1 1 answer 1 2 Info 2 tuples computed Nonetheless these external objects can be accessed f
38. Seot Remarks First line in the answer is the kind of relation view followed by its name in the second line Next and successive pair of lines contain the column name and its type Next lines contain the SQL definition of the view starting with a line containing the delimiter Next lines contain the Datalog definition of the view starting with a line containing the delimiter Finally end of transmission is the last line Both Datalog and SQL outputs are displayed depending on whether pretty print is disabled or not cf Section 5 13 7 i e each statement or rule can be ina single line or multiple lines Fernando Sdenz P rez 159 204 Universidad Complutense de Madrid Datalog Educational System Example Input tapi dbschema v Output Sview v a number integer b string varchar 20 SELECT ALL FROM t NATURAL INNER JOIN s eot e Command tapi is_empty relation _ name Arguments relation_name Relation name either a table or a view which must be enclosed between SQL delimiters if needed Answer Boolean Remarks Return true is relation relation_name is empty i e it contains no tuples in its meaning and false otherwise Example Input tapi is_empty t Output false 5 14 3 TAPI enabled Queries This section shows each supported query for TAPI communication e Query tapi sql_ddl_query Where sql1_dd1_query can be any SQL DDL query cf S
39. Since predicates are defined in general with intensional rules the view p will contain those intensional rules whereas the table will contain the extensional rules facts For instance assuming that the predicate r has been made persisted already in the same connection we assert an intensional rule for p and examine its schema DES gt assert p X r X DES gt dbschema p Info Database Sdes Info View p a number integer Defining SQL statement CREATE VIEW p a AS SELECT ALL FROM p_des_table UNION ALL SELECT ALL rell a FROM r AS rell Datalog equivalent rules p 1 p 2 p X r X Fernando Sdenz P rez 109 204 Universidad Complutense de Madrid Datalog Educational System If you change the current database to the external one and request the schema for p you get DES gt use_db mysql DES gt dbschema p Info Database mysql Info View p a integer 4 which is the schema of view p as provided by the external database system Now the detailed metadata information supplied by des is not available in the external database Also note that the above couple of commands can be simply written as a single one without resorting to change the current database with DES gt dbschema mysql p 5 2 6 Removing Predicate Persistency Finally one can unpersist a given predicate by simply dropping its assertion as in DES gt drop_assertion persistent p a int m
40. Y Info Processing answer X distinct X t X Y answer 1 answer 2 Info 2 tuples computed In addition discarding duplicates can be performed in the context of aggregates DES gt count distinct t X C Info Processing answer C in the program context of the exploded query answer C count p0 X C p0 A distinct t A Fernando S enz P rez 38 204 Universidad Complutense de Madrid Datalog Educational System answer 1 Info 1 tuple computed See also Section 4 1 12 for discarding duplicates in aggregates 4 1 10 Null Values The null value is included in each program signature for denoting unknowns in a similar way it is an inherent part of current relational database systems Comparing null values in Datalog opens a new scenario Two null values are not known to be equal and are not known to be distinct The following illustrates this expected behaviour DES gt null null Info 0 tuples computed DES gt null null Info 0 tuples computed However for the same null value the equality should succeed as in the conjunctive query X null X x A null value is internally represented as NULL ID where ID is a unique identifier an integer Development listings enabled via the command development on allow to inspect these identifiers such as in DES gt development on DES gt p X Y X null Y null X
41. Y Info Processing p X Y ae X SNULL 14 Y SNULL 15 xX Y Info 0 tuples computed DES gt p X Y X null Y null X Y Info Processing p X Y 77 X SNULL 16 Y SNULL 17 X ke Info 0 tuples computed The builtin predicate is_nul1 1 tests whether its single argument is a null value DES gt is_null null Fernando S enz P rez 39 204 Universidad Complutense de Madrid Datalog Educational System is_null null Info 1 tuple computed DES gt X null is_null X Info Processing answer X X null is_null X answer null Info 1 tuple computed Its counterpart predicate is also provided is_not_nu11 1 which is true if its argument is not a null value Note that from a system implementor viewpoint nulls can never unify because they are represented by different ground terms On the other hand disequality is explicitly handled in order to fail when comparing nulls Evaluation of a given expression including at least one null value always returns the same concrete null value Thus two expressions including null values are considered equivalent if they are syntactically equal w r t ground instantiations for null values in particular For instance X null X 1 X 1 succeeds whereas X null Y nul1 X 1 Y 1 and X nu11 X 1 1 X do not 4 1 11 Outer Joins Three outer join operations are provided cf Section 4 5 6 following relational
42. added to the database its head is considered as a query and executed Afterwards the rule is deleted Temporary views are useful for quickly submitting conjunctive queries For instance the view DES gt d X a X not b X computes the set difference between the sets a and b provided they have been already defined Note that the view is evaluated in the context of the program so if you have more rules already defined with the same name and arity of the rule s head the evaluation of the view will return its meaning under the whole set of rules matching the query For instance DES gt a X b X computes the set union of the sets a and b provided they have been already defined 4 1 6 Automatic Temporary Views Automatic temporary views shortly autoviews are temporary views which do not need a head and allows you to write conjunctive queries on the fly When you write a conjunctive query a new temporary relation named answer is built with as many arguments as variables occur in the conjunctive query answer is a reserved word and cannot be used for defining any other relation As an example of an autoview let s consider DES gt a X b Y Info Processing answer X Y a X b Y answer al al answer al bl answer al b2 answer a2 al answer a2 b1 answer a2 b2 answer ai ali answer a3 b1 answer a3 b2 Info 9 tuples computed which computes the Cartesian product of the rel
43. an extension table As far as the query p 1 is subsumed by a previous call p X results in the extension table are reused But if the extension table is cleared then p 1 can be proved DES gt clear_et DES gt p 1 p 1 Info 1 tuple computed Notice that both calls can occur during a computation disabling the opportunity to clear the extension table as in DES gt p X p 1 Info Processing answer X p X p 1 Info 0 tuples computed A similar situation happens with equality Fernando S enz P rez 119 204 Universidad Complutense de Madrid Datalog Educational System DES gt p X X 1 Info Processing answer X p X X 1 Info 0 tuples computed Also notice that if simplification mode is enabled with the command simplification on then this conjunctive query is simplified and computed as follows DES gt p X X 1 Info Processing answer 1 p 1 answer 1 Info 1 tuple computed 5 3 2 Safety for Aggregates and Duplicate Elimination Another source of unsafety departing from the classical notion resides in metapredicates as distinct 2 and aggregates A set variable is any variable occurring in a metapredicate such that it is not bound by the metapredicate For instance Y in the goal distinct X t X Y is a set variable as well as in group_by t X Y X C count Because computing a goal follows SLD order if a set varia
44. and 5 14 3 So feeding unsupported inputs to tapi might produce unexpected results Users of TAPI are expected to ask for other commands and or statements needed for their concrete applications Feedback is welcome 5 14 1 1 Identifiers As SQL identifiers can contain special characters which can be missed with other language constructors they are enclosed between delimiters in such a case This document contains an abbreviated notation name and column_name for table and views in the former and columns in the second When a SQL identifier is written as part of a TAPI input they must be enclosed between the characters L and R left and right delimiters respectively Characters for such delimiters depend on the external Fernando Sdenz P rez 150 204 Universidad Complutense de Madrid Datalog Educational System DBMS For instance MS Access requires and resp but standard SQL defines double quotes for both MS Access does not support this In order to know what are such characters for the current connection one can submit the following commands tapi sql_left_delimiter tapi sql_right_delimiter Datalog identifiers suffer a similar situation but they must be enclosed if needed because containing special characters between single quotes For example tapi listing t Datalog identifiers as returned by DES are not delimited though 5 14 1 2 Kinds of Answers Any input can return either a successful
45. and Local Closed World other nonmonotonic commonsense formalisms including various variants of default reasoning autoepistemic reasoning and other formalisms application specific disambiguation of inconsistent information including defeasible reasoning ConceptBase JJNS98 is a multi user deductive object manager mainly intended for conceptual modeling and coordination in design environments It is multiplatform object oriented it enjoys integrity constraints database updates and several other interesting features The LDL project at MCC lead to the LDL system AOTWZ03 a deductive database system with features such as X Y stratification set and complex terms database updates and aggregates It can be currently used through Internet using a Java enabled client DLV FP96 is a multiplatform system for disjunctive Datalog with constraints true negation la Gelfond amp Lifschitz and queries It includes the K planning system a frontend for abductive diagnosis and Reiter s diagnosis support for inheritance and a SQL front end which prototypes some novel SQL3 features DLVPB is an extension of DLV which provides interfaces with relational databases taking advantage of their efficient implementations to speed up computations XSB RSSWF97 http xsb sourceforge net is an extended Prolog system that can be used for deductive database applications It enjoys a well founded semantics for rules with negative literals in rule bod
46. and the one for Y We do not allow the computation of such rules However if we reorder the two goals as follows less X Y c X Y X lt Y we get the expected result less al b2 less a2 b2 Note then that built in predicates affect declarative semantics i e the intended meaning of the two former views should be the same although actually it is not Declarative semantics is therefore affected by the underlying operational mechanism Notice nonetheless that Datalog is less sensitive to operational issues than Prolog and it could be said to be more declarative First because of terminating issues as already introduced and second because the problematic first view can be automatically transformed into the second computation safe one as we explain next We can check whether a rule is safe in the sense that all its variables are range restricted and then reorder the goals for allowing its computation First we need a notion of safety which intuitively seems clear but that actually is undecidable ZCF 97 Some simple sufficient conditions for the safety of Datalog programs can be imposed which means that rules obeying these conditions can be safely computed although there are rules that even violating some conditions can be actually computed We impose the following weak conditions Ullm95 ZCF 97 for safe rules adapted to our context 1 Any variable X in a ruler is safe if a X occurs in some positive
47. answer al D i answer ai ali answer a2 b1 answer a2 b2 answer ai ali answer ai DI answer ai b i Info 9 tuples computed DES SQL gt Inner Join DES SQL gt select a from a inner join b on a a b b answer a gt answer al Info 1 tuple computed DES SQL gt Left Join DES SQL gt select from a left join b on a a b b answer a a b b gt answer al ali answer a2 null answer a3 null Info 3 tuples computed DES SQL gt Right Join DES SQL gt select from a right join b on a a b b answer a a b b gt answer al ali answer null bl answer null1 b2 Info 3 tuples computed DES SQL gt Full Join DES SQL gt select from a full join b on a a b b answer a a b b gt answer al ali Universidad Complutense de Madrid Datalog Educational System Fernando Sdenz P rez 21 204 Universidad Complutense de Madrid Datalog Educational System answer al null answer a2 null answer a3 null answer null al answer null b1 answer null b2 Info 7 tuples computed DES SQL gt Union DES SQL gt select from a union select from b answer a a gt answer al answer a2 answer a3 answer b1 answer b2 Info 5 tuples computed DES SQL gt Difference DES SQL gt select from a except select from b answer a a gt answer a2 answer a3 Info 2 tuples computed Info Batch file
48. answer with a syntax described for each supported command and statement or an error There are several kinds of answers e Regular o Successful answer with no return data success o Error Serror code text text eot Where code is the error code and text is its textual description which can consist of several lines Last line is the text for denoting end of transmission Error codes are digits starting by either 0 denoting an exception error or 1 denoting a warning or 2 denoting an extended informative message e Boolean Only one line either one of the following o true o false If an error occurs it is output as in the regular answer e Defined specifically for a given command or statement If an error occurs it is output as in the regular answer 5 14 2 TAPI enabled Commands This section shows each supported command for TAPI communication e Command tapi sql_left_delimiter Fernando Sdenz P rez 151 204 Universidad Complutense de Madrid Datalog Educational System Answer Only one line with a single character corresponding to the SQL left delimiter as defined by the database manager either DES or the external DBMS via ODBC Example assuming an ODBC connection to MS Access Input tapi sql_left_delimiter Output Command tapi sql_right_delimiter Answer Only one line with a single character corresponding to the SQL right delimiter as defined by the database man
49. are used to express the negation of a relation either as a query or as a part of a rule body o Disjunctive A disjunctive literal is of the form 1 xr where 1 and r are literals Examples of literals are p r a X not q X b not a b xr a X not q X b 1 lt 2 andX is 1 2 Shorthands for compound goals as not a b are allowed as well which stands for not a b A literal can occur in rule bodies queries and view bodies 4 1 2 Rules Datalog rules have the form head body or simply head Both end with a dot A Datalog head is a positive atom that uses no built in predicate symbol A Datalog body contains a comma separated sequence of literals which may contain built in symbols as listed in Section 4 5 as well as disjunctions 2 4 1 3 Programs DES programs consist of a multiset of rules Programs may contain remarks A single line remark starts with the symbol and ends at the end of line Consulted programs can also contain multi line remarks enclosed between and which can be nested 4 1 4 Queries A positive query is the name of a relation with as many arguments as the arity of the relation a positive literal Each one of these arguments can be a variable or a constant a compound term is not allowed but as an arithmetic expression Built in relations may require relations and conditions as arguments A negative query is written as not Query Queries are typed at the DES system prompt
50. b 0 c 0 b 0 b 0 d 0 b 0 c 0 DES gt strata d 0 1 a 0 2 b 0 1 c 0 1 The first command shows the predicate dependency graph see e g ZCF 97 for the loaded program First nodes in the graph are shown in a list whose elements P are predicates with their arities with the form predicate arity Next arcs in the graph are shown in a list whose elementes are either P Q or P Q where P and Q are nodes in the graph An arc P Q means that there exists a rule such that P is the predicate for its head and Q is the predicate for one of its literals If the literal is negated the arc is negative which is expressed as P Q The graph for this program can be depicted as in Figure 3 a c i b eo a d Figure 3 Predicate Dependency Graph for negation d1l The second command shows the stratum assigned to each predicate This assignment is computed by following an algorithm based on Ullm95 but modified for taking advantage of the predicate dependency graph Strata are shown as a list of pairs PS where P is a predicate and S is its assigned stratum In this example all of the program predicates are in stratum 1 but a which is assigned to stratum 2 This means that if the meaning of a is to be computed then the meanings of predicates in lower strata and only those predicates a depends on have to be firstly computed Since the algorithm strata does not follow a naive bottom up solving only the meanings of required predicates a
51. b int mysql DES gt persistent path a int b int mysql DES gt WITH RECURSIVE path a b AS SELECT FROM edge Fernando Sdenz P rez 114 204 Universidad Complutense de Madrid Datalog Educational System UNION Discarding duplicates ALL is not required SELECT pl a p2 b FROM path pl path p2 WHERE pl b p2 a SELECT FROM path Warning Recursive rule cannot be transferred to external database kept in local database for its processing path_2_1 A B path A C path C B answer path a number integer path b number integer gt answer 1 2 answer 1 3 answer 2 3 Info 3 tuples computed Note the difference against the next query which does not discard duplicates DES gt WITH RECURSIVE path a b AS SELECT FROM edge UNION ALL Keeping duplicates SELECT pl a p2 b FROM path pl path p2 WHERE pl b p2 a SELECT FROM path Warning Recursive rule cannot be transferred to external database kept in local database for its processing path A B path A C path C B answer path a number integer path b number integer gt answer 1 2 answer 1 3 answer 1 3 answer 2 3 Info 4 tuples computed 5 2 9 Caveats 5 2 9 1 Incomplete Meanings If a predicate p which depends on an external relation r is made persistent then it may be the case that the default database engine cannot get the meaning of r but via p as illustrated in the following example
52. database query languages SQL extended relational algebra left right and full outer join Having loaded the example program relop dl we can submit the following queries DES gt c relop DES gt listing a a al a a2 a a3 DES gt listing b b al b b1 b b2 DES gt 135 a X b Y X Y Info Processing answer X Y 1j a X b X Y answer al al answer a2 null answer a3 null Info 3 tuples computed Fernando Sdenz P rez 40 204 Universidad Complutense de Madrid Datalog Educational System DES gt rj a X b Y X Y Info Processing answer X Y rj a X b Y X Y answer al ali answer null bl answer null b2 Info 3 tuples computed DES gt fj a X b Y X Y Info Processing answer X Y fj a X b Y X Y answer al ali answer al null answer a2 null answer a3 null answer null al answer null b1 answer null b2 Info 7 tuples computed Note that the third parameter is the join condition Be aware and do not miss a where condition with a join condition Let s consider the above query lj a X b Y X Y Do not expect the same result as above for the following query DES gt lj a X b X true Info Processing answer X lj a X b X true answer al Info 1 tuple computed Here the same variable X for the relations a and b means that tuples from a and b with the same value are to be j
53. deductive engine So one can extend external database expressiveness with the added features in DES Finally as several ODBC connections are allowed at a time different predicates can be made persistent in different DMBSs which allows for interoperability among external relational engines and the local deductive engine therefore enabling bussiness intelligence applications For instance let s consider MySQL which does not support recursive queries up to its current version 5 5 The following predicate can be made persistent in this RDBMS even when it is recursive DES gt persistent path a int b int mysql DES gt assert path 1 2 DES gt assert path 2 3 DES gt assert path X Y path X Z path Z Y Warning Recursive rule cannot be transferred to external database kept in local database for its processing path X Y path X Z path Z Y DES gt path X Y Fernando Sdenz P rez 113 204 Universidad Complutense de Madrid Datalog Educational System path 1 2 path 1 3 path 2 3 Info 3 tuples computed Here non recursive rules are stored in the external database whereas the recursive one is kept in the local database External rules are processed by MySQL and local rules by the local deductive engine In addition recall that you can use SQL on the current database schema for which the persistent predicate schema is known Then even special SQL features included in DES such a
54. development command in Section 5 13 7 Fernando Sdenz P rez 29 204 Universidad Complutense de Madrid Datalog Educational System Terms Terms can be o Noncompound Variables or constants o Compound As in Prolog they have the form t t1 tn where t is a function symbol functor and ti 1 lt i lt n are terms Up to the current version compound terms can only occur in arithmetic expressions Their function symbols can be any of the built in arithmetic operators and functions cf Section 4 5 2 These operators can be o Infix as addition ee 1 2 o Prefix as bitwise negation e g 1 Examples of terms are r p and p X andX gt Y Atoms An atom has the form a t1 tn where a is a predicate relation symbol and ti 0 lt i lt n are terms If i is 0 then the atom is simply written as a Positive ground atoms are used to build the Herbrand universe There are several built in predicates is for evaluating arithmetical expressions arithmetic functions infix and prefix operators and constants and comparison operators Comparison operators are infix as less than For example 1 lt 2 isa positive atom built from an infix built in comparison operator see Section 4 5 1 Examples of atoms are p r a X 1 lt 2 andX is 1 2 Note that p 1 2 and p t a are not valid atoms Conditions A condition is a Boolean expression containing conjunctions 2 disjunct
55. directory from now on This allows you to run the system whether you have a Prolog interpreter or not in this latter case you have to run the system either on MS Windows Linux or MacOS Although there is no need for further setup and you can go directly to Section 2 2 3 you can also configure a more user friendly way for system start In this way you can follow two routes depending on the operating system 2 2 1 MS Windows 2 2 1 1 Executable Distribution Simply create a shortcut in the desktop for executing the executable of your choice either des exe or deswin exe or des_acide jar The former is a console based executable the second is a windows based executable and the latter is a Java application that includes a call to des exe Executables have been generated with SICStus Prolog so that all SICStus notes in the rest of this document also apply to these executables In addition since it is a portable application it needs to be started from its distribution directory which means that the start up directory of the shortcut must be the distribution directory 2 2 1 2 Source Distribution Perform the following steps 1 Create a shortcut in the desktop for running the Prolog interpreter of your choice 2 Modify the start directory in the Properties dialog box of the shortcut to the installation directory for DES This allows the system to consult the needed files at startup 3 Append the following options to the Pr
56. duplicates as in DES gt assert s X t X DES gt s X s 1 s 1 Info 2 tuples computed In addition recursive rules are duplicate sources as in DES gt assert t X t X DES gt t X t 1 t 1 t 1 t 1 Info 4 tuples computed where two tuples directly come from the two facts for t 1 and the other two from the single recursive rule Again adding the same recursive rule yields DES gt assert t X t X DES gt t X t 1 t 1 t 1 Fernando S enz P rez 36 204 Universidad Complutense de Madrid Datalog Educational System t 1 t 1 t 1 t 1 t 1 t 1 t 1 Info 10 tuples computed where this answer contains the outcome due to two tuples directly from the two facts and four tuples for each recursive rule The first recursive rule is source of four tuples because of the two facts and the two tuples from the second recursive rule Analogously the second recursive rule is source of another four tuples two facts and the two tuples from the first recursive rule The rule of thumb to understand duplicates in recursive rules is to consider all possible computation paths in the dependency graph stopping when a recursive node already used in the computation is reached It is also possible to discard duplicates for an atom with the metapredicate distinct 1 For instance let s consider the following with the same example above DES gt
57. file family sql contains the SQL counterpart code which can be executed with process Famile sol create table father father child insert into father values tom amy insert into father values Jjack fred insert into father values tony carolII insert into father values fred carolIII create table mother mother child insert into mother values grace Tamil insert into mother values amy fred insert into mother values carolI carolII insert into mother values carolII carolIII create view parent parent child as select from father union select from mother create or replace view ancestor ancestor descendant as select parent child from parent union select parent descendant from parent ancestor where parent child ancestor ancestor Fernando Sdenz P rez 174 204 Universidad Complutense de Madrid Datalog Educational System The two example queries above can be formulated in SQL as select from ancestor where ancestor tom select child father mother from father mother where father child mother child And also as RA queries as ra select ancestor tom ancestor project child father mother father zjoin father child mother child mother 6 5 Basic Recursion Problem file recursion dl This example is intended to show that queries involving recursive predicates do terminate thanks to DES fixpoint solving by contrast with Prolog s usual SLD resol
58. from ZCF 97 Fernando S enz P rez 182 204 Universidad Complutense de Madrid Datalog Educational System 6 11 Grammar file grammar d1 Parsers can also be coded as Datalog programs In this example a simple left recursive grammar analyser is coded for the following grammar rules A a A gt Ab A gt Aa It was tested with the input string ababa which is coded with the relation t F T L F for the position of token T that ends at position L t 1 a 2 t 2 b 3 t 3 a 4 t 4 b 5 t 5 a 6 a F L t F a L a F L a F M t M b L a F L a F M t M a L DES gt a 1 6 a 1 6 Info 1 tuple computed 6 12 Fibonacci file fib d1 sql ra The all time classics Fibonacci program can be coded in DES thanks to arithmetic built ins It can be formulated as follows fib 0 1 fib 1 1 fib N F N gt 1 N2 is N 2 fib N2 F2 N1 is N 1 fib N1 F1 F is F2 F1 Since DES is implemented with extension tables computing high Fibonacci numbers is possible with linear complexity DES gt fib 1000 F fib 1000 7033036771142281582183525487718354977018126983635873274 2604905087154537118196933579742249494562611733487750449241765991 16 Taken from FD92 17 Taken from FD92 Fernando Sdenz P rez 183 204 Universidad Complutense de Madrid Datalog Educational System 0881863632654502236471060120533741212738673391111981393731255987 6769009190
59. goal referring to a user defined predicate b r contains some equality goal X Y where Y is safe Y can be a constant which obviously makes X safe c A variable X in the goal X is Expression is safe whenever all variables in Expression are safe 2 A rule is safe if all its variables are safe Fernando Sdenz P rez 117 204 Universidad Complutense de Madrid Datalog Educational System Notice that these conditions currently supported by the system are weak since they assume that user defined predicates are safe which is not always the case but only require analysing locally each rule for deciding weak safety To make these conditions stronger 1 a has to be changed to X occurs in some positive goal referring to a safe user defined predicate and add 3 A predicate is safe if all of its variables are safe The changed conditions would require a global analysis of the program which is not supported by DES up to now The built in predicate is has the same problem as comparison operators as well but it only demands ground its second argument cf condition 1 c above Negation requires its argument to have no unsafe variables In addition to be correctly computed the restrictions in the domains of the safe variables it may contain should be computed before The reader is referred to Section 3 6 in Ullm95 for finding the problems when interpreting rules with negation DES provides a check that allows decidi
60. indexing on extension table on or off resp Default is enabled which shows a noticeable speed up gain in some cases e nospyall Remove all Prolog spy points in the host Prolog interpreter Disable debugging e nospy SPred Arity Remove the spy point on the given predicate in the host Prolog interpreter e spy Pred Arity Set a spy point on the given predicate in the host Prolog interpreter e system Goal Submit Goal to the underlying Prolog system e terminate Terminate the current DES session without halting the host Prolog system Synonym t 5 14 Textual API Rather than providing a Prolog underlying system dependent API DES provides a textual API TAPI Textual Application Programming Interface for its communication to external applications It can used via standard input and output streams as provided by the OS Such interface has been guided by the demands of the ACIDE GUI Graphical User Interface in order to allow users to interact with the system via a Java application This way it is possible to inspect and modify database schema and table contents both those managed by DES and also external data sources as RDBMS s spreadsheets or csv plain files connected by an ODBC connection However this TAPI can be used from any application wrote in any language and running on any platform provided that it can handle input and output standard streams Several existing commands statements and queries can be processed via
61. is assumed to have basic knowledge about Prolog We allow recursive Datalog programs with stratified negation Ullm95 i e normal logic programs without function symbols Stratification Fernando Sd enz P rez 28 204 Universidad Complutense de Madrid Datalog Educational System is imposed to ensure a clear semantics when negation is involved and function symbols are not allowed in order to guarantee termination of queries a natural requirement with respect to a relational database user who is not able to deal with compound data Commands are somewhat different for Prolog programmers as they are accustomed to see Section 5 13 Also exceptions are noted when necessary 4 1 1 Syntax Definitions for Datalog mainly come from the field of Logic Programming Here we follow mainly Lloy87 referring the reader to this book for a more general presentation of Logic Programming Next some definitions for understanding the syntax of programs queries and views are introduced e Numbers Integers and float numbers are allowed A number is a float whenever the number contains a dot between two digits The range depends on the Prolog platform being used Negative numbers are identified by a preceding minus as usual Scientific notation is supported as aEb where a is a fractional number always including a dot and b is an integer which may start with or but it is not required Examples of numbers are 1
62. joins duplicate elimination recursion and grouping with aggregates However there exists an important difference as visibility rules follow SQL instead of RA i e column and relation names are visible to outermost operator applications even when projection or renaming would restrict its visibility With respect to textual syntax we follow Diet01 where arguments of functions are enclosed between parentheses as relations and subscripts and superscripts are delimited between blanks Arguments in infix operators are not enclosed between any delimiters also parentheses can be used to enhance reading Conditions and expressions are built with the same syntax as in SQL Examples below refer to the database defined in examples relop ra 4 3 1 Operators This section includes descriptions for basic additional and extended operators 4 3 1 1 Basic operators Selection o R Select tuples in relation R matching condition 8 Concrete syntax select Condition Relation Example select a lt gt al c Projection 741 an R Return all tuples in R only with columns An An Concrete syntax project Al An Relation Example project b c Set Union Ri U Ro Concrete syntax Relationl1 union Relation2 Example Fernando Sdenz P rez 82 204 Universidad Complutense de Madrid Datalog Educational System a union b Set Difference Ri Ro Concrete syntax Relation1 difference Relation2
63. mutrecursion DES gt consult c des3 0 examples mutrecursion dl This last command assumes that the distribution directory is c des3 0 Synonyms c restore_ddb TAPI enabled e check_db Fernando Sdenz P rez 139 204 Universidad Complutense de Madrid Datalog Educational System Check database consistency w r t declared integrity constraints types existency primary key candidate key foreign key functional dependency and user defined Display a report with the outcome e drop_ic Constraint Drop the specified integrity constraint which starts with and can be either one of e type Table Column Type e nn Table Columns e pk Table Columns e ck Table Columns e k Table Columns RTable RColumns e fd Table Columns DColumns e Goal where Goal specifies a user defined integrity constraint Only one constraint can be dropped at a time Alternative syntax for constraint is also allowed TAPI enabled e listing List the loaded Datalog rules Neither integrity constraints nor SQL views and metadata are displayed e listing Name List the loaded Datalog rules matching Name Neither integrity constraints nor SQL views and metadata are displayed e listing Name Arity List the loaded Datalog rules matching the pattern Name Arity Neither integrity constraints nor SQL views and metadata are displayed e listing Head List the Datalog loaded rules whose heads ar
64. of another relation Next an example of a foreign key assertion is shown DES type p a int type q b int pk q b DES gt fk p a q b However if the relations do not exist an error is raised DES gt fk p a q b Error Relation p has not been typed yet DES gt type p a int type q b int Trying to impose a foreign key with a referenced table which does not have a primary key for matching columns raises an error DES gt fk p a q b Error Referenced column list q b is not a primary key DES gt pk q b DES gt fk p a q b The same constraint cannot be reasserted DES gt fk p a q b Error Trying to reassert an existing constraint DES gt dbschema Info Table s p a number integer FK p a gt q b q b number integer PK b Info No views DES gt assert p 1 Error Foreign key violation p a gt q b when trying to insert p 1 DES gt assert q 1 DES gt assert p 1 DES gt listing p 1 q 1 Info 2 rules listed Several foreign keys may exist for the same relation DES gt type p a int DES gt type q b int DES gt type r a int b int c string DES gt pk p a pk q b DES gt fk r a p a f k r b q b DES gt dbschema r Info Table r a number integer b number integer c string varchar FK r a gt p a FK r b gt q b
65. processor DES gt ra DES RA gt distinct t 3 6 Getting Help You can get useful information with the following commands e help Shows the list of available commands which are explained in Section 5 13 e help Keyword To request help ona given keyword command or built in e builtins Shows the list of built ins which are explained in Section 4 5 Also visit the URL for last information http des sourceforge net Finally you can contact the author via the e mail address fernan sip ucm es 4 Query Languages DES has evolved from a quite simple Datalog interpreter to its current state which relies on a deductive database engine which can be queried with either Datalog SQL or RA languages In addition a Prolog interface is also provided in order to Fernando Sdenz P rez 27 204 Universidad Complutense de Madrid Datalog Educational System highlight the differences between Datalog and Prolog systems Since DES is intended to students it has no full blown features of either state of the art Prolog Datalog or SQL based systems However it has many features that make it appealing as an educational tool along with the novel implementations of declarative debugging sections 5 8 and 5 9 and the test case generator Section 5 10 In this section we describe its four query languages Datalog SQL RA and Prolog The database is shared by all the query languages so that queries or goals can refer to
66. q x select true q p x select true q q x select true p select true p select true q 6 8 Farmer Wolf Goat Cabbage Puzzle file puzzle d1 This example shows the classic Farmer Wolf Goat Cabbage puzzle also Missionaries and Cannibals as another rewritten form The farmer wolf goat and cabbage are all on the north shore of a river and the problem is to transfer them to the south shore The farmer has a boat which he can row taking at most one passenger at a time The goat cannot be left with the wolf unless the farmer is present The cabbage which counts as a passenger cannot be left with the goat unless the farmer is present The following program models the solution to this puzzle The relation state 4 defines the valid states under the specification i e those situations in which there is no danger for any of the characters in our story a state in which the goat is left alone with the cabbage may result in an eaten cabbage and imposes that there is a previous valid state from which we depart from The arguments of this relation are intended to represent from left to right the position north n or south s shore of the farmer wolf goat and cabbage We use the relation safe 4 to verify that a given configuration of positions is valid The relation opp 2 simply states that north is the opposite shore of south and viceversa Initial state state n n n n Farmer takes Wolf state X X U V
67. rule in the in memory database for such a predicate will be persisted too This is to say that for instance if you have persisted already a predicate which is not loaded already and you have a rule asserted a rule for this predicate then the result of restoring its persistency is the union of the asserted rule and the rules in the external database For instance let s consider the following system session DES gt persistent p a int mysql DES gt assert p 1 Now let s assume another system session quit and restart DES Fernando Sdenz P rez 107 204 Universidad Complutense de Madrid Datalog Educational System DES gt assert p 2 DES gt persistent p a int mysql Info Recovering existing data from external database for p DES gt listing p 1 p 2 Info 2 rules listed As it can be seen the resulting database is composed of the union of the external rules and the local rules Finally restoring compiled rules in a different system session does not recover source rules as they were originally asserted They are only recovered as is De compiled form and without textual variable names as they were originally typed in the same system session Let s consider the following DES gt persistent p a int mysql DES gt assert p X X 1 X 2 DES gt listing p X X 1 X 2 Info 1 rule listed DES gt drop_assertion persistent p a int mysql DES gt listing
68. second declaration persists Fernando S enz P rez 48 204 Universidad Complutense de Madrid Datalog Educational System DES gt type p string string DES gt type p int int As well columns can be given names DES gt type p a int b string which is equivalent to the following alternative syntax DES gt type p a int b string However a type declaration for a relation already typed with a different arity is not allowed As will be seen in further sections SQL statements can refer to Datalog relations and SQL does not allow relations of the same name and different arities DES gt type p a int Error Cannot add types to a relation with several arities Relation p A Datalog type declaration is analogous to the creation of a SQL table with the same outcome defining metadata for a relation relation name column names and types DES gt dbschema p Info Table p a number integer b string varchar DES gt drop table p DES gt dbschema p Info No table or view found with that name DES gt create table p a int b string DES gt dbschema p Info Table p a number integer b string varchar It is also possible to omit column names In this case they are automatically provided with names 1 2 and so on DES gt type p int string DES gt dbschema p Info Table p 1 number integer 2 string varchar Let s con
69. shifted left Y places e X gt gt Y Shift right ISO X shifted right places 4 5 4 2 Arithmetic Constants e pi 7 Archimedes constant e e Neperian number Neperian number 4 5 4 3 Arithmetic Functions e sqrt X Square root ISO Square root of X e log X Natural logarithm ISO Logarithm of X in the base of the Neperian number e e 1n X Natural logarithm Synonym for log X Fernando Sdenz P rez 90 204 Universidad Complutense de Madrid Datalog Educational System e log X Y Logarithm Logarithm of Y in the base of X e sin X Sine ISO Sine of X e cos X Cosine ISO Cosine of X e tan X Tangent ISO Tangent of X e cot X Cotangent Cotangent of X e asin X Arc sine Arc sine of X e acos X Arc cosine Arc cosine of X e atan X Arc tangent ISO Arc tangent of X e acot X Arc cotangent Arc cotangent of X e abs X Absolute value ISO Absolute value of X e float X Float value ISO Float equivalent of X if X is an integer otherwise X itself e integer X Integer value Closest integer between X and 0 if X is a float otherwise X itself e sign X Sign ISO Sign of X i e 1 if X is negative 0 if X is zero and 1 if X is positive coerced into the same type as X i e the result is an integer iff X is an integer e gcd X Y Greatest common divisor Greatest common divisor of the two integers X and Y e min X Y Minimum Leas
70. string varchar gt answer al ali answer al b2 answer a2 b i Info 3 tuples computed DES RA gt Left Outer Join DES RA gt a ljoin a a b b b answer a a string varchar b b string varchar gt answer al ali answer a2 null answer a3 null Info 3 tuples computed DES RA gt Right Outer Join DES RA gt a rjoin a a b b b answer a a string varchar b b string varchar gt answer al ali answer null bl answer null b2 Info 3 tuples computed DES RA gt Full Outer Join DES RA gt a fjoin a a b b b answer a a string varchar b b string varchar gt answer al ali answer a2 null answer a3 null Fernando S enz P rez 24 204 Universidad Complutense de Madrid Datalog Educational System answer null bl answer null1 b2 Info 5 tuples computed DES RA gt Union DES RA gt a union b answer a a string varchar gt answer al answer a2 answer a3 answer b1 answer b2 Info 5 tuples computed DES RA gt Difference DES RA gt a difference b answer a a string varchar gt answer a2 answer a3 Info 2 tuples computed DES RA gt Intersection DES RA gt a intersect b answer a a string varchar gt answer al Info 1 tuple computed DES RA gt Grouping DES RA gt group_by a a count true c answer c a string varchar a3 number integer gt answer al 2 an
71. there are none The Cover Texts are certain short passages of text that are listed as Front Cover Texts or Back Cover Texts in the notice that says that the Document is released under this License A Front Cover Text may be at most 5 words and a Back Cover Text may be at most 25 words A Transparent copy of the Document means a machine readable copy represented in a format whose specification is available to the general public that is suitable for revising the document straightforwardly with generic text editors or for images composed of pixels generic paint programs or for drawings some widely available drawing editor and that is suitable for input to text formatters or for automatic translation to a variety of formats suitable for input to text formatters A copy made in an otherwise Transparent file format whose markup or absence of markup has been arranged to thwart or discourage subsequent modification by readers is not Transparent An image format is not Transparent if used for any substantial amount of text A copy that is not Transparent is called Opaque Examples of suitable formats for Transparent copies include plain ASCII without markup Texinfo input format LaTeX input format SGML or XML using a publicly available DTD and standard conforming simple HTML PostScript or PDF designed for human modification Examples of transparent image formats include PNG XCF and JPG Opaque formats include proprietary formats that
72. this interface As well new commands and statements have been added to support the GUI requirements described above Input syntax is as for DES whereas answers follow a concrete format for easing their parsing Any input to this interface must be prepended by the command tapi and cannot be spread beyond a single line as shown next Input tapi test_tapi Output success Notice that after the command tapi another command follows test_tapi which is only intended to test whether a successful connection between the external application and DES can be established If so the answer success is sent to the output stream The usual DES command prompt is not sent as well as no extra blank lines even if compact listings are disabled cf Section 5 13 10 Any input after tapi can also be submitted in the DES command prompt but following the usual DES output instead of the TAPI oriented way Fernando Sdenz P rez 149 204 Universidad Complutense de Madrid Datalog Educational System A typical scenario for accessing DES from an external application is to start a process from this application and connecting adequately input and output streams If run on Windows use the console application des exe for such process otherwise use des both provided in the binary distribution for your concrete operating system 5 14 1 Notes about the Interface e Text in font Courier New are for textual input and output Italized Courier New sta
73. what the individual works permit When the Document is included in an aggregate this License does not apply to the other works in the aggregate which are not themselves derivative works of the Document If the Cover Text requirement of section 3 is applicable to these copies of the Document then if the Document is less than one half of the entire aggregate the Document s Cover Texts may be placed on covers that bracket the Document within the aggregate or the electronic equivalent of covers if the Document is in electronic form Otherwise they must appear on printed covers that bracket the whole aggregate 8 TRANSLATION Translation is considered a kind of modification so you may distribute translations of the Document under the terms of section 4 Replacing Invariant Sections with translations requires special permission from their copyright holders but you may include translations of some or all Invariant Sections in addition to the original versions of these Invariant Sections You may include a translation of this License and all the license notices in the Document and any Warranty Disclaimers provided that you also include the original English version of this License and the original versions of those notices and disclaimers In case of a disagreement between the translation and the original version of this License or a notice or disclaimer the original version will prevail Fernando Sdenz P rez 198 204 A Universidad Compluten
74. 1 1 as understood in extended relational algebra LeftRelationD lt JoinCondition RightRelation e rj LeftRelation RightRelation JoinCondition Right join It stands for the right outer join of the relations LeftRelation and relations RightRelation under the condition JoinCondition expressed as literals cf Section 4 1 1 as understood in extended relational algebra LeftRelation DL JoinCondition RightRelation e f j LeftRelation RightRelation JoinCondition Full join It stands for the full outer join of the relations Left Relation and relations RightRelation under the condition JoinCondition expressed as literals cf Section 4 1 1 as understood in extended relational algebra LeftRelation JoinCondition RightRelation 4 5 7 Datalog Aggregates 4 5 7 1 Aggregate Functions Aggregate functions can only occur in the context of a group_by aggregate predicate see next section and apply to the result set for its input relation e count Variable Return the number of tuples so that the value for Variable is not null e count Return the number of tuples of the result set e sum Variable Return the sum of possible values for Variable ignoring nulls e times Variable Return the product of possible values for Variable ignoring nulls e avg Variable Return the average of possible values for variable ignoring nulls e min Variable Return the minimum value for Variable ignoring nulls e max Variable Re
75. 1 Universidad Complutense de Madrid Datalog Educational System Datalog Educational System V3 0 User s Manual Fernando S enz P rez Grupo de Programaci n Declarativa GPD Departamento de Ingenieria del Software e Inteligencia Artificial DISIA Universidad Complutense de Madrid UCM May 10th 2012 Fernando Sdenz P rez 1 204 Universidad Complutense de Madrid Datalog Educational System Copyright C 2004 2012 Fernando Saenz P rez Permission is granted to copy distribute and or modify this document under the terms of the GNU Free Documentation License Version 1 3 or any later version published by the Free Software Foundation with no Invariant Sections no Front Cover Texts and no Back Cover Texts A copy of the license is included in Appendix A in the section entitled Documentation License Fernando Sdenz P rez 2 204 Universidad Complutense de Madrid Datalog Educational System Contents 1 TANGO CU CHION y ccccssspsdinssteanestniestes totvecucesbecucessesdononteaheobanesteseatueguceatwen venta ER E AE iar 8 1 1 eege Eege 9 Dy CSET ETa E 9 2 Downloading DES irirna wee E E EAE EEE N R e 9 ZEL Source DistibUton EE 10 21 2 EE e GE 10 SE dh E 10 2 1 2 2 DES ACIDE Windows Bune ici casessccasecsnetesniasviandeereridensteceaoeipencanciidies 12 E E E 12 CR E MacOS EE 13 2 2 lnstallinp and Executing eege geed 14 221 NES Mr edu 14 221 1 Executable sbb deser rec 14 Seil
76. 2245245323403501 Info 1 tuple computed Also it is possible to formulate this in SQL even when the next view features non linear recursion file fib sql create view fib n f as select 0 1 union select 1 1 union select fibl nt 1 fibl f fib2 f from fib fibl fib fib2 where fibl n fib2 n 1 and fibl n lt 10 As well next there is a possible RA formulation file ib ra fib n f project 0 1 dual union project 1 1 dual union project fibl nt1 fibl f fib2 f rename fibl nl f1 fib zjoin nl n2 1 and n1 lt 10 rename fib2 n2 2 fib 6 13 Hanoi Towers file hanoi dl Another well known toy puzzle is the towers of Hanoi which can be coded as hanoi 1 A B C hanoi N A B C N gt 1 N1 is N 1 hanoi N1 A C B hanoi N1 C B A We can submit the following query for 10 discs DES gt hanoi 10 a b c hanoi 10 a b c Info 1 tuple computed Note that the answer to this query does not reflect the movements of the discs which can be otherwise shown as the intermediate results kept in the extension table DES gt list_et hanoi Answers hanoi 1 a c b Fernando S enz P rez 184 204 Universidad Complutense de Madrid Datalog Educational System hanoi 1 b a c hanoi 1 c b a hanoi 2 a b c hanoi 2 b c a hanoi 2 c a b hanoi 3 a c b hanoi 3 b a c hanoi 3 c b a hanoi 4 a b c hanoi 4 b c a hanoi 4 c a b hanoi 5 a c b hanoi 5 b a c h
77. 5 16 Notes about the Implementation of DES DES is implemented with the original ideas found in Diet87 TS86 FD92 that deal with termination issues of Prolog programs These ideas have been already used in the deductive database community Our implementation uses extension tables for achieving a top down driven bottom up approach In its current form it can be seen as an extension of the work in Diet87 FD92 in the sense that in addition we deal with negation undefined although incomplete information nulls and aggregates also providing a more efficient tabled mechanism Also the implementation follows a different approach Instead of translating rules we interpret them DES does not pretend to be an efficient system but a system capable of showing the nice aspects of the more powerful form of logic we can find in Datalog systems wrt relational database systems Fernando Sdenz P rez 163 204 Universidad Complutense de Madrid Datalog Educational System 5 16 1 Tabling DES uses an extension table which stores answers to goals previously computed as well as their calls For the ease of the introduction we assume an answer table and a call table to store answers and calls respectively Answers may be positive or negative that is if a call to a positive goal p succeeds then the fact p is added as an answer to the answer table if a negated goal not p succeeds then the fact not p is added Calls are also added t
78. 81 See Section 8 for references to other current deductive database systems 2 Installation 2 1 Downloading DES You can download the system from the DES web page via the URL http des sourceforge net Fernando S enz P rez 9 204 Universidad Complutense de Madrid Datalog Educational System There you can find source distributions for several Prolog interpreters and operating systems and executable distributions for MS Windows Linux and Mac OS 2 1 1 Source Distribution Under the source distribution there are several versions depending on the Prolog interpreter you select to run DES Ciao Prolog BCC97 GNU Prolog Diaz SICStus Prolog SICStus and SWI Prolog Wiele However adapting the code in the file des_glue p1 it could be ported to any other Prolog system See Section 5 16 3 for porting to unsupported systems We have tested DES under several Prolog systems Ciao Prolog 1 14 2 GNU Prolog 1 4 0 SICStus Prolog 4 2 1 and SWI Prolog 6 0 2 and several operating systems MS Windows XP Vista 7 Ubuntu 10 04 1 and MacOSX Snow Leopard The source distribution comes in a single archive file containing the following e readmeDES lt version gt txt A quick installation guide and file release contents e des pl Core of DES including Datalog processor e des deep DCG expansion e des_sql pl SQL processor e des_ra pl RA processor e des_sql_debug pl SQL declarative debugger e des_dl_debug pl Datalog decla
79. ELY NO WARRANTY is x free software and you are welcome to redistribute it x x under certain conditions Type license for details x DE IC DE eta JE JE JE JE JE atta tata tata JE DE HE DE JE DE JE JE JE JE JE JE JE JE E JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE HE XE X X KX K K K KK KX K KX K XK DES gt e deswin exe Windows application executable file as depicted below p 7 BES SICStus 4 2 1 x86 win32 nt 4 Wed Feb 1 01 21 36 WEST 2012 l 2s RRR RRR RRR RRR ERR Ee RR eRe eR ee Oe ee Oe ee OO Ow OO Oe Oo Oe OO OR a DES Datalog Educational System v 3 0 Type help for help about commands amp Fernando Saenz Perez c 2004 2012 GPD DISIA UCM Please send comments questions etc to fernan sip ucm es Web site http des sourceforge net SZ amp t amp This program comes with ABSOLUTELY NO WARRANTY is free software and you are welcome to redistribute it under certain conditions Type license for details KAKAO DESS Please note that the menu bar above is inherited from the host Prolog system and all its settings apply to such system not to DES e d DLL libraries for the runtime system e doc manualDES lt version gt pdf This manual e examples dl Example files which will be discussed in Section 6 Fernando Sdenz P rez 11 204 Universidad Comp
80. ES SQL gt use_db access where access is an example of an already opened connection name 5 1 6 Closing a Connection Closing the current connection is simply done with DES SQL gt close_db You can also specify to close a given connection as in DES SQL gt close_db access 5 1 7 Schema and Data Visibility Any submitted query or command refer to the current connection if not otherwise specified as an argument of a command When opening a connection and automatically making it the current one their data and schema are visible but not the data and schema of other already opened connections In contrast data from the default deductive database are visible for Datalog and RA queries although its schema does not Recall that you can create tables and views in the default database which will be handled by DES but not projected to any external database unless you persist a predicate see Section 5 2 Anyway data from the default deductive database des are not visible for SQL statements for a current connection other than des as they are submitted for processing to the external database In the following system session one creates a table in the default database of DES DDB inserts a value opens a connection and realize that the table schema is not visible but its data do This comes from the fact that first SQL data is translated by DES to Datalog data and second Datalog data can be seamlessly combined with external d
81. Here the system recomputed the strata for the predicate dependency subgraph and informed that it found a stratifiable subprogram for such a query In this simple case no more negations were involved in the subgraph but more elaborated dependencies can be found in other examples cf Sections 6 10 and 6 11 Stratification may be needed for programs without negation as long as a temporary view contains a negated goal Consider the following view under the program relop dl rules in the program with negation are not present in the subgraph for the query d X DES gt d X a X not b X Info Processing d X a X not b X d a2 d a3 Info 2 tuples computed In this view the query d X is solved with a solve by stratum algorithm described in Section 5 16 3 In this case this means that the goal b X is solved before obtaining the meaning of d X because b is in a lower stratum than d and it is needed for the computation of d The basic paradox p not p can be found in the file paradox dl whose model is undefined as you can test with the query p Fernando Sdenz P rez 181 204 Universidad Complutense de Madrid Datalog Educational System 6 10 Parity file parity d1 This example program is intended to compute the parity of a given base relation br X i e it can determine whether the number of elements in the relation cardinality is even or odd by means of the predicates br_is_even and
82. Input considering that table s contains tuples 1 abc null def null null tapi select from s Output answer S a number integer s b string varchar 20 1 abc null def null null Seot Input considering an empty table s tapi select from s Output answer S a number integer s b string varchar 20 Seot 5 15 ISO Escape Character Syntax Special characters in constants and user identifiers can be specified by prepending a backslash to a escape sequence This feature depends on its support by the underlying Prolog system so that the reader is referenced to read corresponding entry in the manual of such system Fernando Sdenz P rez 162 204 Universidad Complutense de Madrid Datalog Educational System Currently escape sequences can only be specified in files to be consulted but not at the command prompt Common escape Sequences are a Alarm ASCII character code 7 b Backspace ASCII character code 8 d Delete ASCII character code 127 e Escape ASCII character code 27 f Form feed ASCII character code 12 n Line feed Newline ASCII character code 10 r Carriage return ASCII character code 13 Go to the start of the line without feeding a new line t Horizontal tab ASCII character code 9 v Vertical tab ASCII character code 11 xhex digit A character code represented by the hexadecimal digits
83. L DQL statements to Datalog clauses on or off resp e show_sql Display whether SQL statements which are sent to an external database are to be displayed e show_sql Switch Enable or disable display of SQL statements which are sent to an external database on or off resp e status Display the current system status i e verbose mode the selected negation algorithm logging elapsed time display program transformation and system version e strata Display the current stratification as a list of pairs PredName Arity Stratum e timing Display whether elapsed time display is enabled e timing Switch Disable or enable either a basic or detailed elapsed time display off on detailed resp e verbose Display whether verbose output is either enabled or disabled on or off resp e verbose Switch Enable or disable verbose output messages on or off resp Fernando Sdenz P rez 146 204 7 Universidad Complutense de Madrid Datalog Educational System 5 13 8 5 13 9 5 13 1 version Display the current DES system version Query Languages datalog Switch to Datalog interpreter all queries are parsed and executed first by Datalog engine If it is not a Datalog query then it is tried first as a SQL statement If it is neither SQL finally it is tried as an RA expression datalog Query Trigger Datalog resolution for the query Query the query is parsed and executed in Datalog but if a parsin
84. L JoinOp Relation or Relation JoinOp Relation JoinCondition Where Relation is as before without any limitation JoinOP is any join operator including INNER JOIN LEFT OUTER JOIN RIGHT OUTER JOIN and FULL OUTER JOIN and JoinCondition can be either ON Condition or USING Column1 ColumnN Where Condition is as described in a WHERE clause and Column ColumnN are common column names of the joined relations Examples Given the tables CREATE TABLE s a int b int CREATE TABLE t a int b int CREATE TABLE v a int b int Fernando Sdenz P rez 68 204 Universidad Complutense de Madrid Datalog Educational System We can submit the following queries SELECT FROM t INNER JOIN s ON t a s a AND t b s b SELECT FROM t NATURAL INNER JOIN s SELECT FROM t INNER JOIN s USING a b SELECT FROM t INNER JOIN s USING a SELECT FROM t INNER JOIN s USING b SELECT FROM t INNER JOIN s ON t a s a AS s v WHERE s a v a SELECT FROM t LEFT JOIN s ON t a s a RIGHT JOIN v ON t a v a SELECT FROM t FULL JOIN s ON t a s a Note The default keyword ALL following SELECT retains duplicates whenever duplicates are enabled command duplicates on In turn DISTINCT discards duplicates But note that if duplicates are disabled both ALL and DISTINCT behave the same i e discarding duplicates 4 2 6 1 1 Top N Queries The number of computed tuples for a select stat
85. N D S D count S gt 1 Info Processing answer D group_by employee N D S D A count S A gt III answer accounting answer sales Info 2 tuples computed Note that the number of employees can also be returned as follows DES gt group_by employee N D S D R count S R gt 1 Info Processing answer D R group_by employee N D S D R count S R gt 1 answer accounting 3 answer sales 3 Info 2 tuples computed Conditions including no aggregates on tuples of the input relation cf SQL FROM clause can also be used cf WHERE conditions in SQL For instance the following query computes the number of employees whose salary is greater than 1 000 DES gt group_by employee N D S S gt 1000 D R count Si Info Processing answer D R in the program context of the exploded query answer D R group_by p2 S D N D R count S p2 S D N employee N D S S gt 1000 answer accounting 2 answer sales 1 Info 2 tuples computed Note that the following query is not equivalent to the former since variables in the input relation are not bound after a grouping computation The following query illustrates this situation which generates a syntax error DES gt group_by employee N D S D R count S S gt 1000 Error Incorrect use of shared set variables in metapredicate Fernando S enz P rez 44 204
86. Prolog 6 0 2 e Changes o License has been relaxed to GNU Lesser General Public License o New versions of command debug_sql1 does not admit a traversing order yet order option removed o Release notes of older DES versions are moved to the new document releasenotesDES pdf e Fixed bugs O Some spanned inputs without leading blanks in multi line mode were not recognised Duplicated object rules were retrieved several times Some commands were not recognized in mixed or uppercase Some listings in development mode did not display all rules Some hypothetical queries led to exceptions Existency of table and attributes in an INSERT SQL statement with a SQL data source was not checked Parsing of a SQL relation separated by a leading space before the comma lead to syntax error Predefined strong constraints relating a tuple of column names were rejected if its lexicographic order did not match the order in which they occur in table definition Running info were logged Fernando S enz P rez 191 204 Universidad Complutense de Madrid Datalog Educational System o Some rules with conjunctions and disjunctions were not parsed correctly from consulted files o GNU Prolog source distribution stopped processing of batch files while encountering a shell command o Predicate dependency graph and strata were not computed after issuing DML SQL statements INSERT DELETE and DQL SQL statement WITH 12 Acknowledgements The a
87. RA formulation pqs x y P union q union project pqs x p y pqs zjoin pqs y p x p union project pqs x q y pqs zjoin pqs y q x q ra select true pqs 6 7 Mutual Recursion files mutrecursion dl sql ra The following program shows a basic example about mutual recursion p a p b q c q d P X q X q X p X Submitting the goal p X we get p a p b p c p d Info 4 tuples computed which is the same set of values for arguments for the query q X The file mrtc dl is a combination of this example and that of the previous section The file mut recursion sql contains the SQL counterpart code which can be executed with process mutrecursion sql sql assert p a assert p b assert q c assert q d Fernando Sdenz P rez 176 204 Universidad Complutense de Madrid Datalog Educational System View q must be given a prototype for view p to be defined create view q x as select from q create or replace view p x as select from q create or replace view q x as select from p Note that it is needed to build a void view for q in order to have it declared when defining the view p The void view is then replaced by its actual definition The contents of both views can be tested to be equal with select from p select from q File mutrecursion ra contains the RA formulation View q must be given a prototype for view p to be defined
88. SQL table and view definitions As rules are not checked for predefined constraints a situations like the following may occur DES gt create table t a int primary key DES gt insert into t values 1 Info 1 tuple inserted DES gt assert t X X 1 DES gt duplicates on DES gt t X t 1 t 1 Info 2 tuples computed Nonetheless if you also want to monitor rules you can otherwise use a user defined constraint such as DES gt create table t a int DES gt insert into t values 1 Info 1 tuple inserted DES gt group_by t X X C count X C gt 1 C gt 1 DES gt assert t X X 1 Error Integrity constraint violation ic X C group_by t X X C count X C gt 1 C gt 1 Offending values in database ic 1 2 Error Asserting rules due to integrity constraint violation 4 2 SQL The syntax recognized by the interpreter is borrowed from the SQL standard This section describes the main limitations features and decisions taken in designing SQL which coexists with Datalog Also we describe the four parts of the supported subset of the SQL language DDL Data Definition Language for defining the database schema DQL Data Query Language for listing contents of the database and DML Data Manipulation Language for inserting and deleting tuples and ISL Information Schema Language Section 4 2 8 resumes the SQL grammar As ODBC connections are allowed some DBMS specific features h
89. The answer to a query is the multi set of atoms matching the query which are deduced in the context of the program from both the extensional and intensional database A query with variables for all the arguments of the queried relation gives the whole set of deduced facts meaning defining the relation as the query a X in the example of Section 3 If a query contains a constant in an argument position it means that the query processing will select the facts from the meaning of the relation such that the argument position matches with the constant i e analogous to a select relational operation This is the case of the query a a3 in the same example You can also write conjunctive queries on the fly such as a X b X see Section 4 1 6 Built in comparison operators listed in Section 4 5 1 can be safely used in queries whenever their arguments are ground at evaluation time excepting equality which performs unification Disjunctive queries are also allowed too such as a X b X Concluding a query follows the same syntax as rule bodies Fernando S enz P rez 31 204 Universidad Complutense de Madrid Datalog Educational System If only a limited number of tuples in the answer are required one can submit the query as top N Query where N is the maximum number of tuples to be returned 4 1 5 Temporary Views Temporary views allow you to write conjunctive queries on the fly A temporary view is a rule which is
90. Windows XP Vista 7 32 bit with both SICStus Prolog and SWI Prolog executables and sources IBM DB2 v9 7 200 358 Oracle Database Express Edition 11g Release 2 also tested with Windows 7 64 bit and SWI Prolog 6 0 0 64 bit SQL Server Express 2008 including spatial components MySQL 5 5 9 PostgreSQL 9 1 3 Access 2003 Excel 2003 CSV text files 5 2 Persistency Since DES 3 0 it is possible to make predicates persist on either an external database or datasheet or text file i e any data source supported by an ODBC connection This sections describes how to persist a predicate use it examine its schema unpersist it and also lists a couple of caveats 5 2 1 Persisting a Predicate An assertion is used to declare a persisted predicate as in DES gt persistent p a int mysql where its first argument is the predicate and its schema and the second one is the ODBC connection name This name can be omitted if the current connection is the one you want to use to persist the predicate as in DES gt current_db Info Current database is mysql DBMS mysql DES gt persistent p a int You can confirm that predicate p has been declared as persistent with DES gt list_persistent Fernando Sdenz P rez 104 204 Universidad Complutense de Madrid Datalog Educational System mysql p a number integer where the connection name is shown followed by a semicolon and the predic
91. _sql pl and des_ra pl contain the SQL and RA processor respectively Files des_sql_debug pl and des_dl_debug pl contain the SQL and Datalog declarative debuggers File des_types p1 contains the type checking and inference system File des_tc pl contains the SQL test case generator code The last file des_glue pl contains Prolog system specific code which vary from a system to another Adapting the predicates found there should not pose problems provided that the Prolog interpreter and operating system feature some basic characteristics mainly about the file system commands In particular finite domain constraints is a must for supporting several features of DES such as type inference and test case generation If you plan to port DES to other systems not described here you will have to modify the system specific Prolog file to suit your system If so and if you want to figure as one of the system contributors please send an e mail message with the code and reference information to fernan sip ucm es accepting that your contribution will be under the GNU Lesser General Public License See the appendix for details 5 16 5 Differences among Platforms Ciao SWI and SICStus Prolog implementations use a sort which eliminates duplicates whereas GNU Prolog implementation does not In its current version the Ciao system forces to use some directives for using several basic Prolog primitives This can only be done by writing them in the core file
92. ad Complutense de Madrid Datalog Educational System DES gt type q a int DES gt select from p q where p ax lt q a answer p a number integer q a number integer gt answer 1 2 Info 1 tuple computed DES gt p zjoin p a lt q a q answer p a number integer q a number integer gt answer 1 2 Info 1 tuple computed And persistent predicates can be combined even with external data coming from other ODBC connection as in DES gt open_db access DES gt dbschema t Info Database access Info Table t a INTEGER 4 DES gt p X t X Info Processing answer X p X t X answer 1 Info 1 tuple computed Here the current database is access and all its data is available as already introduced in Section 5 1 2 in particular the table t which contains in particular the tuple t 1 As well one can retract the rules previously asserted For instance DES gt retract p 1 DES gt retract p X r X 5 2 3 Processing a Persistency Assertion Processing a persistency assertion means to make persistent a predicate i e all of its current rules as well as rules added afterwards are stored in a persistent media as a relational database A fact is projected to a table whereas a rule is translated into a SQL view Each persisted predicate is translated into a table for holding such facts and a view which is the union of all the SQL translations for its rules Tr
93. age symbols start comments User identifiers must start with a letter and consist of letters and numbers otherwise a user identifier can be enclosed between quotation marks both square brackets and double quotes are supported and contain any characters Next SQLstmt stands for a valid SQL statement SQLstmt DDLstmt DMLstmt DQOLstmt ISLstmt LEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEES DDL Data Definition Language statements LEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEES DDLstmt CREATE OR REPLACE TABLE CompleteConstrainedSchema CREATE OR REPLACE TABLE TableName LIKE TableName CREATE OR REPLACE VIEW ViewSchema AS DQLstmt Fernando Sdenz P rez 75 204 Universidad Complutense de Madrid Datalog Educational System RENAME TABLE TableName TO TableName ae VIEW ViewName TO ViewName Ge TABLE IF EXISTS TableName TableName Extended syntax following MySQL 5 6 Ge VIEW ViewName DROP DATABASE Schema RelationName RelationName Att Att CompleteConstrainedSchema RelationName Att Type ColumnConstraint ColumnConstraint Att Type ColumnConstraint ColumnConstraint TableConstraints CompleteSchema RelationName Att Type Att Type Type CHAR n fixed length string of n characters CHARACTER n equivalent to the former CHAR fixed length string of 1 character VARCHAR n variable length st
94. ager either DES or the external DBMS via ODBC Example assuming an ODBC connection to MS Access Input tapi sql_right_delimiter Output Command tapi Zed Answer Only one line with the full path DES was started from Example Input tapi cd Output c des Command tapi cd Path Answer Only one line with the full new path Example Input tapi cd examples Output c des examples Fernando Sdenz P rez 152 204 Universidad Complutense de Madrid Datalog Educational System e Command tapi consult File tapi c File tapi File Answer Information about the loaded program and a final line containing eot Examples Input tapi family Output Info 11 rules consulted Seot Input tapi c family fact Output Warning N gt 0 may raise a computing exception if non ground at run time Warning N1 is N 1 may raise a computing exception if non ground at run time Warning F is N Fl may raise a computing exception if non ground at run time Warning Next rule is unsafe because of variable s F N fac N F N gt 0 N1 is N 1 fac N1 F1 F is N Fl Info 13 rules consulted eot e Command tapi reconsult Files tapi r Files tapi Files Answer Information about the loaded program and a final line containing eot Example Input tapi family Output Info 11 rules consulted Seot e Command Fernando Sdenz P
95. aggregates etc Let s consider the following session in which we are interested in specifying a directed tree a connected graph with no cycles DES gt verbose on Info Verbose output is on DES gt consult paths Info Consulting paths edge a b edge a c edge b a edge b d path X Y path X Z edge Z Y path X Y edge X Y end_of_file Info 6 rules consulted Info Computing predicate dependency graph Info Computing strata DES gt path X X Info Parsing query Info Constraint successfully parsed Info Checking user defined integrity constraint over database path X X Info Computing predicate dependency graph Info Computing strata Error Integrity constraint violation ic X path X X Offending values in database ic b ic a Info Constraint has not been asserted The constraint path X X specifies that a path from a node to itself is not allowed As the consulted program contains a cycle involving nodes a and b the constraint is violated and therefore it is not asserted Offending values are listed in this case all the values involved in any cycle you can try out other edges and see the outcome Another use is to first specify the constraint and then a graph However don t be tempted to submit the constraint and consult the program the constraint will be removed since consulting a program amounts to erase the existing database including user define
96. ame Set the current log to the given filename and mode write overwrite existing file if any or creates a new one or append append to the contents of the existing file nolog Disable logging Informative apropos Keyword Display detailed help about Keyword which can be a command or built in Synonyms help Ferna ndo Sd enz P rez 143 204 Universidad Complutense de Madrid Datalog Educational System builtins List predefined operators functions and predicates check Display whether integrity constraint checking is enabled compact_listings Display whether compact listings are enabled dbschema Display the database schema Tables views and constraints dbschema Name Display the database schema for the given connection view or table name TAPI enabled dbschema Connection Name Display the database schema for the given view or table name in the given connection dependent_relations Relation Display the name of relations that directly depend on relation Relation Arity TAPI enabled dependent_relations Relation Arity Display in format Name Arity those relations that directly depend on relation Relation Arity TAPI enabled development Display whether development listings are enabled development Switch Enable or disable development listings on or of resp These listings show the source to source translations needed to handle null values Datalog outer join built ins
97. and Applications Symposium International 0 344 353 2005 H Garcia Molina J D Ullman J Widom Database Systems The Complete Book Prentice Hall 2002 M A W Houtsma and P M G Apers Algebraic optimization of recursive queries Data amp Knowledge Engineering Volume 7 Issue 4 March 1992 Fernando Sd enz P rez 202 204 Universidad Complutense de Madrid Datalog Educational System IRIS2008 JGJ 95 KLW95 KSSD94 KT81 Lloy87 Mink87 MN82 MS11 PDR91 Robi65 RS09 RSSS94 RSSWF97 RU95 IRIS Reasoner http iris reasoner org M Jarke R Gallersd rfer M A Jeusfeld M Staudt S Eherer ConceptBase a deductive object base for meta data management In Journal of Intelligent Information Systems Special Issue on Advances in Deductive Object Oriented Databases Vol 4 No 2 167 192 1995 System available at http www i5 informatik rwth aachen de CBdoc M Kifer G Lausen J Wu Logical Foundations of Object Oriented and Frame Based Languages Journal of the ACM vol 42 p 741 843 1995 W Kiessling H Schmidt W Strauss and G D nzinger DECLARE and SDS Early Efforts to Commercialize Deductive Database Technology VLDB Journal 3 pp 211 243 1994 C Kellogg and L Travis Reasoning with Data in a Deductively Augmented Data Management System H Gallaire J Minker and J Nicolas eds Advances in Data
98. and disjunctive literals duplicates Display whether duplicates are enabled hypothetical Display whether hypothetical queries are enabled on or not off sql_left_delimiter Display the SQL left delimiter as defined by the current database manager either DES or the external DBMS via ODBC TAPI enabled sql_right_delimiter Display the SQL left delimiter as defined by the current database manager either DES or the external DBMS via ODBC TAPI enabled help Display resumed help on commands Shorthands h Fernando Sdenz P rez 144 204 Universidad Complutense de Madrid Datalog Educational System e help Keyword Display detailed help about Keyword which can be a command or built in Synonyms apropos e is_empty relation_name Display St rue if the given relation is empty and false otherwise TAPI enabled e list_tables List table names TAPI enabled e list_table_ schemas List table schemas TAPI enabled e list_table_ constraints table_name List table constraints for table_name TAPI enabled e list_views List view names TAPI enabled e list_view_schemas List view schemas TAPI enabled e negation Display the selected algorithm for solving negation strata or et_not e pdg Display the current predicate dependency graph e pdg PredName Display the current predicate dependency graph restricted to the first predicate found with name PredName e pdg PredName Arity Disp
99. and not computed if unsafety or uncomputability is detected and cannot be repaired because program transformation is disabled or there is no way Notice that there can be unsafe or uncomputable rules already consulted than can yield an incorrect result or raise a run time exception Fernando Sdenz P rez 118 204 Universidad Complutense de Madrid Datalog Educational System Concluding one can expect a correct answer whenever no unsafe uncomputable rule has been asserted to an empty database Recall that the local analysis relies on the weak condition that assumes that the consulted rules are safe Next an example of unsafe rule including negation is provided As introduced such a rule when asserted raises an error but it is asserted in any case in order to show its misbehaviour DES gt assert q 0 DES gt assert p X not q X Error not q X might not be correctly computed because of the unrestricted variable s X Warning This rule is unsafe because of variable s X DES gt p X Info 0 tuples computed As the domain of X in p X is not range restricted no tuples are found in the left to right top down search If we submit a query as p 1 the negation not q 1 should be proven DES gt p 1 Info 0 tuples computed However as illustrated there is no tuples in the answer for such a query The misbehaviour of the rule for p 1 emerges here due to the way answers are computed via
100. anda Garc a Ruiz and Fernando S enz P rez Date 5 2007 Description Tool for the declarative debugging of Datalog programs License LGPL Contact Yolanda Garc a Ruiz Implementor e ACIDE A Configurable Development Environment Authors Diego Cardiel Freire Juan Jos Ortiz S nchez Delf n Rup rez Ca as SI 2006 2007 Miguel Mart n L zaro SI 2007 2008 and Javier Salcedo G mez SI 2010 2011 leaded by Fernando S enz Date 3 2007 ACIDE 0 1 first version 11 2008 ACIDE 0 7 current alpha version Description This project is aimed to provide a multiplatform configurable integrated development environment which can be configured in order to be used with any development system such as interpreters compilers and database systems Features of this system include project management multifile editing syntax colouring and parsing on the fly which informs of syntax errors when editing programs prior to the compilation License GPL Project Web Page http acide sourceforge net e Emacs development environment Author Markus Triska Date 2 22 2007 Description Provides an integration of DES into Emacs Once a Datalog file has been opened you can consult it by pressing F1 and submit queries and commands from Emacs This works at least in combination with SWI Prolog it depends on the s switch other systems may require slight modifications License GPL Project Web Page http stud4 tuwien ac at e0225855 index h
101. anoi 5 c b a hanoi 6 a b c hanoi 6 b c a hanoi 6 c a b hanoi 7 a c b hanoi 7 b a c hanoi 7 c b a hanoi 8 a b c hanoi 8 b c a hanoi 8 c a b hanoi 9 a c b hanoi 9 c b a hanoi 10 a b c Info 27 tuples in the answer set 6 14 Other Examples Directory examples include some other examples as the files bom d1 bill of materials and trains d1 train connections which show more example applications including negation Other examples are orbits d1 a cosmos tiny database sg dl same generation for a family database Ce dl transitive closure and empTraining ra sql taken from Diet01 Also the folder persistent contains examples for persisting predicates the folder ontology includes examples of authoring ontologies including some documentation and folders DLDebugger and SQLDebugger include examples for debugging Datalog programs and SQL views respectively 7 Contributions This section collects the contributions from external developers up to now e Test Case Generator Authors Rafael Caballero Roldan Yolanda Garcia Ruiz and Fernando Sdenz P rez Date 10 2009 upgraded version supported since DES 1 8 0 Description Tool for generating test cases for SQL views License LGPL Contact Yolanda Garcia Ruiz Implementor Fernando Sdenz P rez 185 204 Universidad Complutense de Madrid Datalog Educational System e Datalog Declarative Debugger Authors Rafael Caballero Rold n Yol
102. anslating rules into SQL views includes an adaptation of Draxler s Prolog to SQL compiler Drax92 Any rule belonging to the definition of a predicate pred which is being made persistent is expected in general to involve calls to other predicates Each callee such other called predicate can be Ss An existing relation in the external database Fernando Sdenz P rez 106 204 Universidad Complutense de Madrid Datalog Educational System An already persisted predicate which is loaded in the local database An already persisted predicate which is no yet loaded in the local database A predicate which has not been made persistent yet For the first two cases besides making pred persistent nothing else is performed when processing its persistency assertion For the third case a persistent predicate is automatically restored in the local database c f next section i e it is made available to the deductive engine For the fourth case each non persistent predicate is automatically made persistent if possible inferring its types This is needed in order for the external database to be aware of a predicate which is only known by the deductive engine so far as this database will eventually compute pred However not all rules can be made persistent for a number of reasons including that the external database does not support some features and the translations of some built ins are not supported yet In the curre
103. any object defined using any language However there are some dependent issues that must be taken into account For instance once a Datalog fact is loaded into the database the relation it defines can be queried in Datalog But if one wants to access this relation from either SQL or RA two alternatives are provided 1 Define the same relation in SQL via a create table statement Section 4 2 4 1 and 2 Declare types for the table Section 4 1 14 1 This particular issue comes from the fact that Datalog relations have unnamed attributes and a positional reference is used for accessing those relations In turn SQL and RA use a notational syntax giving names to relation arguments To illustrate the first alternative let s consider the following session DES gt assert t 1 DES gt t X t 1 Info 1 tuple computed DES gt select from t Error Unknown table or view t DES gt create table t a int DES gt select from t answer t a gt answer 1 Info 1 tuple computed The error above reflects that t is not a known object in the database scheme Following the second alternative to access a Datalog relation from SQL DES gt assert t 1 DES gt type t a int DES gt select from t answer t a gt answer 1 Info 1 tuple computed LI Datalog Since Datalog stems from Prolog we have adopted almost all the Prolog syntax conventions for writing Datalog programs the reader
104. as not Goal instead of the usual Goal When a goal is solved instead of displaying the variable substitution for the answer the goal is displayed with the substitution applied as in DES Prolog gt CO t 1 type for more solutions lt Intro gt to continue t 2 type for more solutions lt Intro gt to continue no AS Built ins Most built ins are shared by the four languages For instance w r t comparison operators the only difference is the less or equal lt operator used in Datalog and Prolog This operator is different from the used in SQL and RA which is written as lt The former is written that way since in Prolog and Datalog it is distinguished from the implication to the left operator lt SQL does not provide implications so the SQL syntax seems to be more appealing since the order of the two symbols matches the order of words Arithmetic expressions are constructed with the same built ins in the three languages However in Datalog and Prolog you need to use the infix is cf Section 4 5 2 The built in predicates is_nul1 1 and is_not_nu11 1 belong to the Datalog language Also consult Section 5 3 for limitations regarding safety in the use of built ins in Datalog 4 5 1 Comparison Operators All comparison operators are infix and apply to terms For the inequality and disequality operators greater than less than etc numbers are compared in terms of their arithmet
105. as the dot in Datalog and the semicolon in SQL and RA But when writing a long query as usual in SQL breaking down the sentence along several lines enhances readability This is also possible in DES by enabling multi line mode with the command multiline on However in this scenario the terminating character must be issued in order to know when to finish parsing the input query Returning to single line mode is just by issuing multiline off With multi line input multi line remarks enclosed between and are also allowed Note that nested remarks are supported too as First remark Second nested remark S E 5 6 Development Mode This section is focused at those interested in modifying and extending the system So from a system implementor viewpoint it is handy to show several implementation specific issues such as source to source transformations and internal representation of null values To this end the command development on off has been made available Let s consider the following system session DES gt development off DES gt assert p X X 1 X 2 DES gt assert c C count p X X C DES gt assert q 1 DES gt assert 1 X Y 13 p X q Y X Y DES gt listing c C count p X X C 1 X Y 1j p X q X Y p X X se L X 2 q 1 Info 4 rules listed DES gt 1 X Y 1 1 1 1 2 null Fernando Sdenz P rez 122 204 Univ
106. atabases EDB DES gt create table t a int Create table t in DDB DES gt insert into t values 1 Insert t 1 in DDB Info 1 tuple inserted DES gt select from t Select data from DDB answer t a number integer gt answer 1 Info 1 tuple computed DES gt open_db mysql Open an EDB Fernando Sdenz P rez 99 204 Universidad Complutense de Madrid Datalog Educational System DES gt select from t Select data from EDB Error ODBC Code 1146 As t is not defined in EDB MySQL ODBC 5 1 Driver mysqld then error 5 5 9 Table test t doesn t exist DES gt t X Predicate t is known to DDB and can be queried from Datalog t 1 Info 1 tuple computed In this way you can also combine data from DES and the external data source Next system session example shows this by creating a new table in the external database and combining above predicate t 1 defined in DDB with a new table s created in EDB DES gt create table s a int Create table s in EDB DES gt insert into s values 2 Insert s 2 in EDB Info 1 tuple inserted DES gt select from s Select data from EDB answer a integer 4 gt Note the different type w r t DDB answer 2 Info 1 tuple computed DES gt t X s Y Join t 1 DDB with s 1 EDB Info Processing answer X Y t X s Y answer 1 2 Info 1 tuple computed 5 1 8 Integrity Constraints ODBC C
107. ate schema Also if you have type information declared already you can simply refer to the predicate with its name and arity in the persistency assertion DES gt Zuse db Sdes DES gt create table p a int DES gt use_db mysql DES gt persistent p 1 DES gt list_persistent mysql p a number integer The general form of a persistency assertion is as follows persistent PredSpec Connection This assertion makes a predicate to persist on an external RDBMS via an ODBC connection PredSpec can be either the pattern PredName Arity or PredName Schema where Schema can be either ArgNamel ArgNameN or ArgNamel Typel ArgNameN TypeN If a connection name is not provided the current open database is used The local default database des cannot be used to persist but an ODBC connection 5 2 2 Using Persistent Predicates You can assert facts as usual and query the persisted predicate p 1 as the following example shows DES gt assert p 1 DES gt p X p 1 Info 1 tuple computed And as expected it can seamlessly be combined with other non persistent predicates as in DES gt assert q 2 DES gt p X q Y X lt Y Info Processing answer X Y p X q Y X lt Y answer 1 2 Info 1 tuple computed where q 2 is in the meaning of q 1 Also you can use SQL or RA languages to query such persistent predicates as in Fernando Sdenz P rez 105 204 Universid
108. ated with the SQL statement SQOLStatement as its definition Note that column names are mandatory Examples DES gt dbschema Info Table s s a number integer b number integer PK a b u b number integer c number integer PK b t a number integer b number integer c number integer d number integer PK a c FK t b d gt s a b FK t b gt u b Info View s v a number integer b number integer c number integer d number integer Defining SQL Statement SELECT ALL FROM t WHERE a gt 1 Datalog equivalent rules v A B C D t A B C D A gt 1 w a number integer b number integer Defining SQL Statement SELECT ALL t a s b FROM t s WHERE t a gt s a Datalog equivalent rules w A B t A C D E s F B A gt F Info No integrity constraints Note that primary key constraints follow the table schema and inferred types are in the view schema 4 2 4 3 Dropping Tables DROP TABLE IF EXISTS TableName TableName This statement drops the table schema corresponding to each one of the provided names TableName deleting all of its tuples whether they were inserted Fernando Sd enz P rez 64 204 Universidad Complutense de Madrid Datalog Educational System with INSERT or with the command assert and rules which might have been added via assert If the optional clause IF EXISTS is included d
109. ates and duplicate elimination e Non linear recursive queries e Recursive queries are not restricted w r t aggregates or nested computations as usual RDBMS s are IBM DB2 MS SQL Server SUN Oracle MySQL e Simplified recursive queries are allowed Although supported there is no need for using a WITH clause e Hypothetical queries which are a novel proposal out of the standard e Set operators build relations which can be used wherever a data source is expected FROM clause e Null values are supported along with outer joins full left and right e Aggregate functions allowed in expressions at the projection list and HAVING conditions GROUP BY clauses are also allowed e View support Any relation built with a SQL query can be defined as a view even recursive queries e Supported database integrity constraints include type constraints existency nullability primary keys candidate keys and referential integrity constraints e Parentheses can be used elsewhere they are needed and also for easing the reading of statements e Suggestions are provided for misspelled table view and column names when similar entries are found Fernando Sdenz P rez 60 204 Universidad Complutense de Madrid Datalog Educational System 4 2 3 Datalog vs SQL With respect to Datalog some decisions have been taken e As in Datalog user identifiers are case sensitive table and attribute names This is not the no
110. ations a and b provided they have been already defined as Fernando Sdenz P rez 32 204 Universidad Complutense de Madrid Datalog Educational System a al a a2 a a3 b bl b b2 b al 4 1 7 Underscored Variables An underscored variable a variable starting with the underscore symbol _ is handled similar to Prolog It is assumed to be of no interest for the answer so that they are discarded from the answer should they occur in the body of a query view or autoview even in its head For instance computing the projection of a relation t with respect to its first argument can be simply done as follows DES gt assert t 1 2 DES gt assert t 2 3 DES gt t X _ Info Processing answer X t X _ answer 1 answer 2 Info 2 tuples computed instead of having to resort to an autoview such as DES gt p X t X Y Info Processing p X t X Y p 1 p 2 Info 2 tuples computed Also let s consider other situation as follows DES gt duplicates off DES gt t X Y t 1 1 t 1 2 t 3 3 Info 3 tuples computed DES gt t X X t 1 1 t 3 3 Fernando Sdenz P rez 33 204 Universidad Complutense de Madrid Datalog Educational System Info 2 tuples computed If you use instead underscored variables you get one answer tuple DES gt t _X _X Info Processing answer t _X _X answer Info 1 tuple co
111. ave been added as well as non standard features in ISL 4 2 1 Main Limitations e The projection list consists of column references column table column alias column wildcards table alias alias references arithmetic expressions and SQL statements Other expressions might be supported in further releases Fernando S enz P rez 59 204 Universidad Complutense de Madrid Datalog Educational System e A limited coverage of database integrity constraints e Strong typing Different numeric type values cannot be compared eg real and integer Also there is no provision for automatic type casting e No provision for ordering results order by clause e No insertions deletions updates into views e Limited syntax error reports The parser does not inform about all the possible syntax error causes but for table view and column misspelled names However syntax errors from ODBC connections are displayed 4 2 2 Main Features As main features we highlight e Data query data definition and data manipulation language parts provided e Subqueries nested queries without depth limits e Correlated queries tables and relations in nested subqueries can be referenced by the host query For example SELECT FROM t SELECT a FROM s WHERE t a s a e Subqueries in comparisons as SELECT a FROM t WHERE t a gt SELECT a FROM s e Table relation and expression aliases with full scope e Support for duplic
112. ble is used after the metapredicate as in distinct X t X Y GO then this is an unsafe goal as in the call to distinct variable Y is not bound and all tuples in t 2 are considered for computing its outcome Swapping both subgoals yields a safe goal So data providers for set variables are only allowed before their use in such metapredicates Along compilations unsafe rules can be automatically generated as in the translations of outer joins However they are safe because of their use unsafe arguments of such rules are always given as input in goals So mode information for predicates is handled throughout program compilations to detect truly unsafe rules avoiding to raise warnings about system generated rules Notice however that you can still manually write an unsafe call to these system generated predicates yielding to incorrect results as the following examples illustrates DES gt assert t 1 DES gt assert s 2 DES gt assert 1 X 1j t X s Y X Y DES gt development on DES gt listing p0 X Y p1 X Y p0 X NULL A t X Fernando S enz P rez 120 204 Universidad Complutense de Madrid Datalog Educational System not p1 X Y p1 X Y X Y t X s Y 1 X 1j p0 X Y s 2 t 1 Info 6 rules listed DES gt p0 X Y p0 1 NULL 0 Info 1 tuple computed DES gt list et Answers not p1 1 A t 1
113. br_is_odd respectively The predicate next defines an ascending chain of elements in br based on their textual ordering where the first link of the chain connects the distinguished node nil to the first element in br The predicates even and odd define the even resp odd elements in the chain The predicate has_preceding defines the elements in br such that there are previous elements to a given one the first element in the chain has no preceding elements The rule defining this predicate includes an intended error fourth rule in the example which will be used in Section 6 13 to show how it is caught by the declarative debugger Pairs of non consecutive elements in br between X Z br X br br Z X lt Y Y lt Z Consecutive elements in the sequence starting at nil next X Y br X br X lt Y not between X Y next nil X br X not has_preceding X Values having preceding values in the sequence has_preceding X br X br Y Y gt X terror Y gt X should be Y lt xX Values in an even position of the sequence including nil even nil even Y odd X next X Y Values in an odd position of the sequence odd Y even X next X Y Succeeds if the cardinality of the sequence is even br_is_even even X not next X Y Succeeds if the cardinality of the sequence is odd br_is_odd odd X not next X Y Base relation br a br b 15 Adapted
114. can be read and edited only by proprietary word processors SGML or XML for which the DTD and or processing tools are not generally available and the machine generated HTML PostScript or PDF produced by some word processors for output purposes only Fernando Sdenz P rez 194 204 Universidad Complutense de Madrid Datalog Educational System The Title Page means for a printed book the title page itself plus such following pages as are needed to hold legibly the material this License requires to appear in the title page For works in formats which do not have any title page as such Title Page means the text near the most prominent appearance of the work s title preceding the beginning of the body of the text The publisher means any person or entity that distributes copies of the Document to the public A section Entitled XYZ means a named subunit of the Document whose title either is precisely XYZ or contains XYZ in parentheses following text that translates XYZ in another language Here XYZ stands for a specific section name mentioned below such as Acknowledgements Dedications Endorsements or History To Preserve the Title of such a section when you modify the Document means that it remains a section Entitled XYZ according to this definition The Document may include Warranty Disclaimers next to the notice which states that this License applies to the Document These Warranty Disclaimers are considere
115. candidate key foreign key s a t a functional dependency a b and user defined integrity constraint t X s X X a gt t a An D An nm D UH a gt b t X s X X eot Command tapi relation_schema relation_name Arguments relation_name Relation name either a table or view which must be enclosed between SQL delimiters if needed Answer relation_kind relation_name column_name type column_name type column_name type Seot Remarks Return relation schema of relation_name First line in the answer is the kind of relation either table for a table or view for a view followed by its name in the second line Next and successive pair of lines contain the column name and column type Example Input tapi relation_schema t Output Stable t a Fernando Sdenz P rez 158 204 Universidad Complutense de Madrid Datalog Educational System number integer Seot e Command tapi drop_ic constraint Arguments constraint Constraint following Datalog syntax cf Section 4 1 14 8 Answer Regular Example Input tapi drop_ic pk s b Output Ssuccess e Command tapi dbschema view_name Arguments view_name View name as a SQL identifier which needs to be enclosed between SQL delimiters if needed Answer relation_kind relation_name column_name type column_name type SQL SQL Datalog Datalog
116. ces from the validity of others As an example we show a debugger session for the query br_is_even in the program parity d1 which has been changed to contain an error in the following rule has_preceding X br X br Y Y gt X Serror Y gt X should be Y lt X In this case the user expects the answer for the query br_is_even to be br As even because the relation br contains two elements a and b However the answer returned by the system is which means that the corresponding query was unsuccessful The available command for starting a debugging session is debug_datalog Goal where Goal is a basic goal i e no conjunctive or disjunctive goals are allowed Therefore the user can start a typical debugging session as follows DES gt debug_datalog br_is_even Debugger started Is br b br b valid v non valid n v v Is has_preceding b valid v non valid n v n Is br X br b br a valid v non valid n v v Error in relation has_preceding 1 Witness query has_preceding b In this particular case only three questions are necessary to find out that the relation has_preceding is incorrectly defined Fernando Sdenz P rez 128 204 Universidad Complutense de Madrid Datalog Educational System 5 9 SQL Declarative Debugger As in the previous section here we focus on a declarative approach to debugging following CGS12a former version of the debugger is based on CGS11b
117. cessing answer D R group_by employee N D S D R count answer accounting 3 answer null 2 answer resources 1 answer sales 5 Info 4 tuples computed Note that two employees are not assigned to any department yet nolan and norton This query behaves as a SQL user would expect though nulls do not have to represent the same data value in spite of this such tuples are collected in the same bag If we rather want to count active employees those with assigned salaries we pose the following query DES gt group_by employee N D S D R count S Info Processing answer D R group_by employee N D S D R count S answer accounting 3 answer null1 0 answer resources 1 answer sales 3 Info 4 tuples computed Note that null departments have no employee with assigned salary Counting the number of departments from the relation employee needs to discard duplicates as in DES gt count_distinct employee N D S D T Info Processing answer T count_distinct employee N D S D T Fernando Sdenz P rez 43 204 Universidad Complutense de Madrid Datalog Educational System answer 3 Info 1 tuple computed Conditions including aggregates on groups can be stated as well cf having conditions in SQL For instance the following query counts the active employees of departments with more than one employee DES gt group_by employee
118. cludes all their local rules including those which are the result of SQL compilations and external rules persisted predicates Their integrity constraints and SQL table and view definitions are removed The extension table is cleared and the predicate dependency graph and strata are recomputed e abolish Name Arity Delete the predicates matching the pattern Name Arity This includes all their local rules including those which are the result of SQL compilations and external rules persisted predicates Their integrity constraints and SQL table and view definitions are removed The extension table is cleared and the predicate dependency graph and strata are recomputed e assert Head Body Add a Datalog rule If Body is not specified it is simply a fact Rule order is irrelevant for Datalog computation The extension table is cleared and the predicate dependency graph and strata are recomputed e consult FileName Load the Datalog program found in the file Filename discarding the rules already loaded integrity constraints and SQL table and view definitions The extension table is cleared and the predicate dependency graph and strata are recomputed The default extension d1 for Datalog programs can be omitted Examples Assuming we are on the distribution directory we can write DES gt consult examples mutrecursion which behaves the same as the following DES gt consult examples mutrecursion dl DES gt consult examples
119. computes the number of active employees by department DES gt c D C count employee N D S S C Info Processing c D C count employee N D S S D C Fernando Sdenz P rez 45 204 Universidad Complutense de Madrid Datalog Educational System c accounting 3 ec null 0 c resources 1 c sales 3 Info 4 tuples computed Note that the system adds to the aggregate predicate an argument with the list of grouping variables which are the ones occurring in the first argument of the aggregate predicate that also occur in the head This code translation is required for the aggregate predicate to be compute although such form has not been made available to the user Having conditions are also allowed including them as another goal of the first argument of the aggregate predicate as for instance in the following view which computes the number of employees that earn more than the average DES gt count employee N D S avg employee N1 D1 S1 S1 A S gt A C Info Processing answer C in the program context of the exploded query answer C count p2 A S D N C p2 A S D N employee N D S avg employee N1 D1 S1 S1 A S gt A answer 2 Info 1 tuple computed Note that this query uses different variables in the same argument positions for the two occurrences of the relation employee Compare this to the following query which computes the number of employ
120. constraints which are called user defined integrity constraints from now on All of them can be declared and the system monitors their fulfilment which is the default behaviour However the command check off allows to disable constraint checking All predefined integrity constraints apply to facts but type constraints which also apply to rules Also user defined constraints apply to facts and rules A comma separated sequence of predefined integrity constraints is allowed to specify multiple constraints in a single input 4 1 14 1 Type A type constraint specifies the values in a domain a predicate argument table column in relational jargon may take An example of type constraint declaration at the command prompt is as follows DES gt type p int string This is equivalent to the following alternative syntax DES gt type p int string Allowed types include the following where each row in the first column contains type synonyms varchar String of unbounded length string char N SEIN String with length up toN char String with length 1 integer i Integer number int float Real number real Precision and range depend on the underlying Prolog system Subsequent type declarations are allowed for the same predicate and arity the last declaration is the one to persist overriding previous type declarations for such predicate The following session is possible and thus the
121. d duplicate elimination integrity constraints ODBC connections to external relational database management systems RDBMSs Datalog and SQL tracers a textual API for external applications and novel approaches to hypothetical SQL queries declarative debugging of Datalog queries and SQL views test case generation for SQL views null values support tabled outer join and aggregate predicates The system is implemented on top of Prolog and it can be used from a Prolog interpreter running on any OS supported by such interpreter Moreover Windows Linux and MacOSX executables are also provided We have developed DES aiming to have a simple interactive multiplatform and affordable system not necessarily efficient for students so that they can get the fundamental concepts behind a deductive database with Datalog Relational Algrebra and SQL as query languages SQL is supported with a reasonable coverage of the standard for teaching purposes Supported extended relational algebra includes duplicates outer joins and recursion Other deductive systems are not fully suited to our needs due to the absence of some characteristics DES does offer for our educational purposes This system is not targeted as a complete deductive database so that it does not provide transactions security and other features present in current database systems There are two relevant enhancements in the current release Predicate persistency supported by external databa
122. d integrity constraints Instead use the reconsult command DES gt verbose on Info Verbose output is on DES gt cd examples Info Current directory is c fernan research bddeduc des des3 0 examples DES gt path X X Info Parsing query Info Constraint successfully parsed Info Checking user defined integrity constraint over database Fernando Sdenz P rez 56 204 Universidad Complutense de Madrid Datalog Educational System path X X Info Computing predicate dependency graph Warning Undefined predicate s path 2 Info Computing strata DES gt reconsult paths Info Consulting paths edge a b edge a c edge b a edge b d Info Checking user defined integrity constraint over database path X X Info Computing predicate dependency graph Info Computing strata path X Y path X Z edge Z Y Info Checking user defined integrity constraint over database path X X Info Computing predicate dependency graph Info Computing strata Error Integrity constraint violation ic X path X X Offending values in database ic b ic a path X Y edge X Y File c fernan research bddeduc des des3 0 examples paths dl Lines 10 10 end ot file Info 5 rules consulted Info Computing predicate dependency graph Info Computing strata Note that the first rule for path is not rejected since in the already consulted program it is s
123. d students that have passed a basic level course level 0 View intensive defines as intensive students those in the table allInOneCourse together with the students that have completed the three initial levels However this view definition is erroneous We have forgotten to check that the courses have been completed flag pass Finally the main Fernando Sdenz P rez 132 204 Universidad Complutense de Madrid Datalog Educational System view awards selects the students in the basic but not in the intensive courses Suppose that we try the query select from awards and that in the result we notice that the student Anna is missing We know that Anna completed the basic course and that although she registered in the three initial levels she did not complete one of them and hence she is not an intensive student Thus the result obtained by this query is nonvalid So the user starts the debugger as Anna is not among the possibly large list of student names produced by view awards The debugging session proceeds as follows DES gt process examples SQLDebugger awards1 DES gt debug_sql awards Info Debugging view awards 1 awards Carla Input Is this the expected answer for view awards y n m mT w wN a h n m Anna Info Debugging view intensive Input Should intensive include a tuple of the form Anna y n a Iyl n Info Debugging view standard Input Should standard include a tuple
124. d to be included by reference in this License but only as regards disclaiming warranties any other implication that these Warranty Disclaimers may have is void and has no effect on the meaning of this License 2 VERBATIM COPYING You may copy and distribute the Document in any medium either commercially or noncommercially provided that this License the copyright notices and the license notice saying this License applies to the Document are reproduced in all copies and that you add no other conditions whatsoever to those of this License You may not use technical measures to obstruct or control the reading or further copying of the copies you make or distribute However you may accept compensation in exchange for copies If you distribute a large enough number of copies you must also follow the conditions in section 3 You may also lend copies under the same conditions stated above and you may publicly display copies 3 COPYING IN QUANTITY If you publish printed copies or copies in media that commonly have printed covers of the Document numbering more than 100 and the Document s license notice requires Cover Texts you must enclose the copies in covers that carry clearly and legibly all these Cover Texts Front Cover Texts on the front cover and Back Cover Texts on the back cover Both covers must also clearly and legibly identify you as the publisher of these copies The front cover must present the full title with all words of
125. database are not checked for primary key and foreign key constraints Next a very simple example is reproduced to illustrate basic constraint handling DES SQL gt create or replace table u b int primary key c int DES SQL gt create or replace table s a int b int primary key a b DES SQL gt create or replace table t a int b int c int d int primary key a c foreign key b d references s a b foreign key b references u b DES SQL gt insert into t values 1 2 3 4 Error Foreign key violation t b d gt s a b when trying to insert t 1 2 3 4 Info 0 tuples inserted DES SQL gt insert into s values 2 4 Info 1 tuple inserted DES SQL gt insert into t values 1 2 3 4 Error Foreign key violation t b gt u b when trying to insert t 1 2 3 4 Info 0 tuples inserted DES SQL gt insert into u values 2 2 Info 1 tuple inserted DES SQL gt insert into t values 1 2 3 4 Info 1 tuple inserted DES SQL gt listing s 2 4 t 1 2 3 4 u 2 2 4 2 4 2 Creating Views CREATE OR REPLACE VIEW ViewName Columnl ColummN AS SQLStatement Fernando Sdenz P rez 63 204 Universidad Complutense de Madrid Datalog Educational System This statement defines the view schema in a similar way as defining tables If the optional clause OR REPLACE is used the view is dropped if existed already Other tuples or rules asserted with the command assert are not deleted The view is cre
126. des DES gt persistent p a int mysql DES gt assert p 1 DES gt show_compilations on DES gt select from p Info SQL statement compiled to answer A p A answer p a number integer gt answer 1 Info 1 tuple computed DES gt use_db mysql DES gt select from p answer a integer 4 gt answer 1 Info 1 tuple computed Note that in the first case first SELECT above when the current database is des DES solves the query in this case retrieving tuples from DDB and in the second case second SELECT above the query is directly submitted to the EDB which solves it In the first case the SQL statement is compiled to Datalog and solved by the deductive engine and in the second one data and metadata are collected from EDB and shown as a result Retrieved types from an external database differ in general to those managed by DES as it can be seen in this example This is not an issue as long as equivalent types are found in this case number integer is considered as equivalent to integer 4 as numeric size constraints are not handled by DES up to now As already introduced in Section 5 1 7 even when a connection is opened their data and metadata is not known unless it becomes the current database as illustrated next DES gt use_db mysql DES gt create table q a int DES gt insert into q values 2 Info 1 tuple inserted DES gt select from q answer a integer 4 gt
127. distinct t X Info Processing answer X distinct t X answer 1 Info 1 tuple computed Such query is equivalent to the following SQL statement provided that metadata is available for the relation t DES gt type t a int DES gt select distinct from t answer t a gt answer 1 Info 1 tuple computed As it would be expected duplicates are only discarded for the call distinct Atom but not for other occurrences of Atom during query solving Thus DES gt t X distinct t X Info Processing answer X t X distinct t X answer 1 answer 1 answer 1 answer 1 Fernando S enz P rez 37 204 Universidad Complutense de Madrid Datalog Educational System answer 1 answer 1 answer 1 answer 1 answer 1 answer 1 Info 10 tuples computed Compare this to the call DES gt t X t X Info Processing answer X t X t X answer 1 answer 1 Info 100 tuples computed A subset of arguments in an atom can be selected for discarding duplicates To this end the metapredicate distinct 2 is provided Its first argument is the list of variables for which duplicates are not required i e each concrete assignment of values to all variables in the list must be different So let s consider the following session DES gt listing t 1 1 t 1 2 t 2 1 Info 3 rules listed DES gt distinct X t X
128. dscsterndvecheslsversstivae teenie 75 42 8 SOL Grammars ee es i ts ee ee en a C R 75 4 3 Extended Relational Al eebra ic ides ncsiclec tas ctisnconsdnver atieaythedinvevrtoens teens 82 43 1 CMP CLAL OLS mensage e ound ld achat E alee stata aes 82 43 119 BASIC OPO LALONS sernir iee t ren r tiene ee AnA PERA a E een EERE np PETs PEER 82 4 3 1 2 ee 83 Ce E ER En 84 43 2 Rec rtsionin eer 85 43 3 RA EE 85 AAS Prolog isa chteeon cach deceice ices a e a a a E n a i ES 87 D iene oee R E a E E E a S 87 eh Comparis n Operators ries arr n E E o S 87 4 5 2 Datalog and Prolog Arithmetic ett 88 4 5 3 E Erres sss a ae A iets Sie Hee ee 89 45 4 Arithmetic RI E 89 EN e Operators csi sierveeniaierecduerecevennutasnanrminlceccndioncccvamciiinet 90 ASAD Arithmetic Coristants s nosists aei aai 90 ABAD Arithmetic Functions ys e er A E E 90 TE E e eE EE EEEE E EE eh as E E a EE RE 92 45 6 D talog Outer Joins E 92 Fernando S enz P rez 4 204 Universidad Complutense de Madrid Datalog Educational System 4 5 7 y Datalog Aggregates piini isein eain Sn Ya eair E EEA vances hi does tee dolls 92 457 1 Aggregate e sisser r eek i E eaen EEEE E ED SEa EEEE 92 45 72 Group by Predicate E 92 4 5 7 Aggregate Predicates ieira i E EEA irate 93 4 5 8 Datalog Null related Predicates ergeet geeeegee get eptnentectiodingsiteeeres 93 459 opreegend eer ee 93 45 10 Ee 94 RE 94 5 1 RDBMS connections via ODBC eene 94 5 1 1 Opening a
129. e 2 Further Investigations of Deduction in Relational Databases H Gallaire and J Minker eds Logic and Databases Plenum Press 1978 D Diaz http www gnu org software prolog S W Dietrich Extension Tables Memo Relations in Logic Programming IV IEEE Symposium on Logic Programming 1987 S W Dietrich Understanding Relational Database Query Languages Prentice Hall 2001 M Derr S Morishita and G Phipps Design and Implementation of the Glue NAIL Database System In Proc of the ACM SIGMOD International Conference on Management of Data pp 147 167 1993 Draxler Chr A Powerful Prolog to SQL Compiler CIS Bericht 92 61 Centrum f r Informations und Sprachverarbeitung Ludwig Maximilians Universit t M nchen 1992 C Fan and S W Dietrich Extension Table Built ins for Prolog Software Practice and Experience Vol 22 7 pp 573 597 July 1992 R Fikes P J Hayes and I Horrocks OWL QL a language for deductive query answering on the Semantic Web J Web Sem 2 1 19 29 2004 Wolfgang Faber and Gerald Pfeifer DLV homepage since 1996 url http www dlvsystem com C C Green and B Raphael The Use of Theorem Proving Techniques in Question Answering Systems Proceedings of the 234 ACM National Conference Washington D C 1968 S Greco I Trubitsyna and E Zumpano NP Datalog A Logic Language for NP Search and Optimization Queries Database Engineering
130. e SQL statements in relop sql you can type abolish and process relop sql Note that the extension can be omitted in the process command Here we depart from the Datalog interpreter and if you are to submit SQL queries it is useful to switch to the SQL interpreter via the command sql as inputs will be parsed only by the SQL parser Otherwise it will be tried to be identified as a Datalog input and then as a SQL input Note that in the file relop sql1 listed below strings are enclosed between apostrophes This is not needed in the Datalog language In order to execute the contents of this file type process relop sql Switch to SQL interpreter sql Creating tables create or replace table a a create or replace table b b create or replace table c a b Listing the database schema dbschema Inserting values into tables insert into a values al insert into a values a2 insert into a values a3 insert into b values bl insert into b values b2 insert into b values al insert into c values al b2 insert into c values al al insert into c values a2 b2 Testing the just inserted values select from a select from b select from c Projection select a from c Selection select a from a where a a2 Cartesian product select from a b Fernando Sdenz P rez 168 204 7 Universidad Complutense de Madrid Datalog Educational System Inner Join
131. e covers all the loaded rules for t 1 LO t a Info 2 rules listed Should any other constraint remains asserted other than a type constraint a type constraint cannot be changed DES gt type p a int b string Error Cannot change type assertion while other constraints remain 4 1 14 1 1 Types on Intensional Database Types can also be declared for predicates of the intensional database i e those predicates defined at least with rules not only with facts So asserting a new type constraint over an intensional relation will trigger type checking inferring types along Fernando Sdenz P rez 50 204 Universidad Complutense de Madrid Datalog Educational System the predicate dependency graph restricted to the typed predicate Let s consider the following situation as an example DES gt listing s a t 1 t X s X Info 3 rules listed DES gt type t int Error No type tuple covers all the loaded rules for t 1 t 1 t X s X Info 2 rules listed 4 1 14 1 2 Types on Propositional Relations Finally propositional relations are also subject of beign typed of course with an empty list of arguments DES gt type a DES gt dbschema a Info Table a The alternative syntax becomes shorter in this case indeed DES gt type a 4 1 14 2 Nullability Existency Constraint Columns can be imposed to contain a concrete value rather than a null The next sy
132. e initial query Q w r t the program P This graph represents how the meanings of queries are constructed See more details in CGS07 The second phase consists of traversing the CG to find either a buggy vertex or a set of related incorrect vertices The vertex associated to the initial query Q is marked automatically as non valid by the debugger The rest of the vertices are marked initially as unknown In order to minimize the number of questions asked by a declarative debugger several traversing strategies have been studied Caba05 Silv07 However these strategies are only adequate for declarative debuggers based on trees and not on graphs The currently implemented strategy already contains some ideas of how to minimize the number of questions in a CG e First the debugger asks about the validity of vertices that are not part of cycles in order to find a buggy vertex if it exists Only when this is no longer possible the vertices that are part of cycles are visited e Each time the user indicates that a vertex Query FactSet is valid i e the validity of the answer for the subquery Query is ensured the tool changes to valid all the vertices with queries subsumed by Query e Each time the user indicates that a vertex Query FactSet is non valid the tool changes to non valid all the vertices with queries subsumed by Query The last two items help to reduce the number of questions deducing automatically the validity of some verti
133. e line mode which the dot is optional Rules in a consulted file may span on multiple lines and ending dot is mandatory irrespective the multi line mode SQL o User identifiers including tables views column names are case sensitive o Some incorrect SQL statements are not rejected as those containing a GROUP BY clause and columns in the projection list which do not occur in the grouping list Rather they raise exceptions at run time o Computable SQL statements follow the grammar in Section 4 2 8 of this manual The current grammar parses extra clauses which cannot be computed yet e g ORDER BY ANY o See also Section 5 1 7 regarding ODBC connections SQL debugger o SQL debugging is not supported for ODBC connections up to now Test case generator o Source distribution for Ciao partially supports this feature o Source distribution for GNU Prolog does not support negative integers o Test case generation is not supported for ODBC connections up to now SQL tracer o SQL tracing is not supported for ODBC connections up to now Miscellanea o Enabling duplicates can notably harm performance cf Fibonacci example o Users should not write predicate identifiers starting with the symbol Otherwise unexpected behaviour might happen o Batch processing cannot be nested Prolog systems specific issues O Safety checks for aggregates and distinct 2 are not supported in Ciao source version Fernando S enz P rez 189 204
134. e subsumed by the head Head Neither integrity constraints nor SQL views and metadata are displayed e listing Head Body List the Datalog loaded rules that are subsumed by Head Body Neither integrity constraints nor SQL views and metadata are displayed e reconsult FileName Load a Datalog program found in the file Filename keeping the rules already loaded The extension table is cleared and the predicate dependency graph and strata are recomputed TAPI enabled See also consult Filename Synonyms r e restore_ddb Filename Restore the Datalog database in the given file same as consult Constraints type nullability primary key candidate key functional dependency foreign key and user defined are also restored if present in Filename e retract Head Body Delete the first Datalog rule that unifies with Head Body or simply with Head if Body is not specified In this case only facts are deleted The extension Fernando Sdenz P rez 140 204 7 Universidad Complutense de Madrid Datalog Educational System 5 13 2 5 13 3 table is cleared and the predicate dependency graph and strata are recomputed retractall Head Delete all the Datalog rules whose heads unify with Head The extension table is cleared and the predicate dependency graph and strata are recomputed save_ddb force Filename Save the current Datalog database to the file Filename If option force is included no question is asked to
135. e te de te de te Se de tee dete de te dee te te te te de de te te te Se te ke ete te tte tee te te aOR tO a te DES gt examples aggregates dl Grammar bytes Lexicon Configuration des 1 1 NumLines 98 INS 00 29 46 2 1 2 3 Linux From the same URL above you can download a Linux executable distribution in a single archive file containing the following e readmeDES lt version gt A quick installation guide and file release contents e des Console executable file e doc manualDES lt version gt pdf This manual e examples dl Example files which will be discussed in Section 6 e license license A verbatim copy of the GNU Public License for this distribution The following screenshot has been taken in Ubuntu 10 04 1 Fernando Sdenz P rez 12 204 Universidad Complutense de Madrid fernan fernan ubuntu mnt DES DES3 0 Archivo Editar Ver Terminal Ayuda LESSEE SESS SESE SSS SESS SESS ESSE ESSE SESSA EE E EE EE E EE KEE EE EE DES Datalog Educational System v 3 0 Type help for help about commands Fernando Saenz Perez c 2004 2012 GPD DISIA UCM Please send comments questions etc to fernan sip ucm es Web site http des sourceforge net This program comes with ABSOLUTELY NO WARRANTY is free software and you are welcome to redistribute it under certain conditions Type license for details LESSEE ESSE SESS SSS SSS SSS SSS EE EE EE EE EE EE EE E EE E EE EE E EE EE EE EE
136. e used in projection lists and conditions 4 5 4 Arithmetic Built ins This section contains the listings for the supported arithmetic operators constants and functions Fernando S enz P rez 89 204 Universidad Complutense de Madrid Datalog Educational System 4 5 4 1 Arithmetic Operators The following operators are the only ones allowed in arithmetic expressions where X and Y stand also for arithmetic expressions e VS Bitwise negation ISO Bitwise negation of the integer X e X Negative value ISO Negative value of its single argument X e xX Y Power ISO X raised to the power of Y e xX Y Power Synonym for X Y e Xx Y Multiplication ISO X multiplied by Y e X Y Real division ISO Float quotient of X and Y e X Y Addition ISO Sum of Xand Y e X Y Subtraction ISO Difference of X and Y e X Y Integer quotient ISO Integer quotient of X and Y The result is always truncated towards zero e X rem Y Integer remainder ISO The value is the integer remainder after dividing X by Y i e integer X integer Y X Y The sign of a nonzero remainder will thus be the same as that of the dividend e X Y Bitwise disjunction ISO Bitwise disjunction of the integers X and Y e X Y Bitwise conjunction ISO Bitwise disjunction of the integers X and Y e X Y Bitwise exclusive or Bitwise exclusive or of the integers X and Y e X lt lt Y Shift left ISO X
137. ected in the result set Then we can debug that view as follows DES gt debug_sql Guest Info Debugging view Guest 1 Guest 1 Mark Costas 2 Guest 2 Helen Kaye 3 Guest 3 Robin Scott Input Is this the expected answer for view Guest y n m mT w wN a h n n Info Debugging view CatsAndDogsOwner 1 CatsAndDogsOwner 1 Wilma 2 CatsAndDogsOwner 2 Lucky 3 CatsAndDogsOwner 3 Rocky Fernando Sdenz P rez 129 204 Universidad Complutense de Madrid Datalog Educational System Input Is this the expected answer for view CatsAndDogsOwner y n m mT w wN a h y n Info Debugging view NoCommonName NoCommonName 1 NoCommonName 2 NoCommonName 3 LA M l Input Is this the expected answer for view NoCommonName y n m mT w wN a h y n Info Debugging view LessThan6 LessThan6 1 LessThan6 2 LessThan6 3 LessThan6 4 amp WDN Ee Input Is this the expected answer for view LessThan6 y n m mT w wN a h y y Info Debugging view AnimalOwner AnimalOwner 1 Kitty cat AnimalOwner 1 Wilma dog AnimalOwner 2 Lucky dog AnimalOwner 2 Wilma cat AnimalOwner 3 Oreo cat AnimalOwner 3 Rocky dog AnimalOwner 4 Cecile turtle AnimalOwner 4 Chelsea dog D JO DG M Input Is
138. ection 4 2 4 Answer Regular Examples Fernando Sdenz P rez 160 204 Universidad Complutense de Madrid Datalog Educational System Input tapi create table t a int Output Ssuccess Input tapi rename table t to q Output Ssuccess e Query tapi sql_dml_query Where sq1_dm1_query can be any SQL DML query cf Section 4 2 5 Answer If successful one single line with the number of affected tuples Examples Input tapi insert into t values 3 Output 1 Input tapi insert into t values 3 Output Serror 0 Type mismatch number integer table declaration Seot e Query tapi sql_dql_query Where sq1_dq1_query can be any SQL DQL query cf Section 4 2 6 Answer relation_name column_name type column_name type value value Fernando S enz P rez 161 204 Universidad Complutense de Madrid Datalog Educational System value value Seot Where relation_name is the name of the answer relation column_name is a column name type is the column type value is the column value is the record delimiter and eot is the end of the transmission Remarks This DOL statement returns in the first line the name of the answer relation the first column name and its type in the next two lines and so for all of its columns Then each or the tuples in the relation preceded by the record delimiter Last line is the end of transmission Examples
139. ed receipt of a copy of some or all of the same material does not give you any rights to use it 10 FUTURE REVISIONS OF THIS LICENSE The Free Software Foundation may publish new revised versions of the GNU Free Documentation License from time to time Such new versions will be similar in spirit to the present version but may differ in detail to address new problems or concerns See http www gnu org copyleft Each version of the License is given a distinguishing version number If the Document specifies that a particular numbered version of this License or any later version applies to it you have the option of following the terms and conditions either of that specified version or of any later version that has been published not as a draft by the Free Software Foundation If the Document does not specify a version number of this License you may choose any version ever published not as a draft by the Free Software Foundation If the Document specifies that a proxy can decide which future versions of this License can be used that proxy s public statement of acceptance of a version permanently authorizes you to choose that version for the Document 11 RELICENSING Massive Multiauthor Collaboration Site or MMC Site means any World Wide Web server that publishes copyrightable works and also provides prominent facilities for anybody to edit those works A public wiki that anybody can edit is an example of such a server A Massive M
140. ed to the corresponding tables add option or the contents of the tables replaced by the generated test case tuples replace option For experimenting with the domain of attributes we provide the command tc_domain Min Max which defines de range of values the integer attributes may take This range is determinant in the search of test cases in a constraint network that That is executing the view using as input data for the tables those in the PTC Fernando Sdenz P rez 135 204 Universidad Complutense de Madrid Datalog Educational System can easily become too complex as long as involved views grow So keeping this domain small allows to manage bigger problems String constants occurring in all the views on which the view for the test case generated depends are mapped to integers in the same domain starting from 0 So the size of the domain has to be larger enough to hold at least the string constants in those views Also we provide the command tc_size Min Max for specifying the size of the test case generated in number of tuples Again keeping this value small helps in being able to cope with bigger problems Currently we provide support for integer and string attributes Binary distributions and both SICStus and SWI Prolog source distributions allow the functionality described GNU Prolog source distribution only allows non negative integers in the domain declaration Ciao Prolog source distribution partiall
141. ed with the command save_ddb Filename which saves in a plain file the Datalog rules in memory Later they can be restored with restore_ddb Filename this command is only an alias for consult In the following session the current database is stored abolished cleared and finally restored All the data including the ones interactively added have been recovered DES gt save_ddb db dl DES gt abolish DES gt restore_ddb db dl Info 19 rules consulted DES gt a X a al a a2 a a3 a a4 Fernando Sdenz P rez 18 204 Universidad Complutense de Madrid Datalog Educational System Info 4 tuples computed Another useful command is list_et which lists in particular the answers already computed Following the last series of queries and commands above we submit Answers a al a a2 a a3 a a4 Info 4 tuples in the answer table Calls a A Info 1 tuple in the call table Here we can see that the computed meaning of the queried relation is stored in an extension table as well as the last call cf sections 5 16 1 and 5 16 2 Unless either the database is changed e g via assert or retract commands or a temporary view see Section 4 1 6 executed or the command clear_et is submitted the extension table keeps computed results otherwise it is cleared 3 2 SQL Mode In this mode queries are sent to the SQL processor whereas commands cf Section 5 13 are se
142. ees so that each one of them earns more than the average salary of his corresponding department Here the same variable name D has been used to refer to the department for which the counting and average are computed DES gt count employee N D S avg employee N1 D S1 S1 A S gt A C Info Processing answer C in the program context of the exploded query answer C count p2 A S N C p2 A S N employee N D S avg employee N1 D S1 S1 A S gt A answer 3 Info 1 tuple computed Fernando S enz P rez 46 204 Universidad Complutense de Madrid Datalog Educational System Also as a restriction of the current implementation keep in mind that having conditions including aggregates as the one including the average computations above can only occur in the first argument of an aggregate The following query which should be equivalent to the last one would generate a run time exception DES gt v D avg employee N1 D S1 S1 A count employee N D S S gt A C Error S gt A will raise a computing exception at run time Warning This view is unsafe because of variable s A Finally recall that expressions including aggregate functions are not allowed in conjunction with aggregate predicates but only in the context of a group_by predicate 4 1 13 Disjunctive Bodies As introduced in Section 4 1 1 rule bodies can contain disjunctions such as the one conta
143. ements can be limited with the so called Top N queries ISO 2008 includes this as a final clause in the select statement SELECT DISTINCT ALL ProjectionList FROM Rels FETCH FIRST Integer ROWS ONLY However DES also provides another non standard but common form in other RDBMS s of such queries SELECT TOP Integer DISTINCT ALL ProjectionList You can switch the order of the top and distinct clauses and even specify both forms of Top N queries in the same statement as long as they express the same limit 4 2 6 1 2 The dual table The dual table is a special one row one column table present by default in all Oracle database installations It is suitable for use in selecting a pseudocolumn with no data source As propositional relations are also allowed in DES dual does not need a Fernando Sdenz P rez 69 204 Universidad Complutense de Madrid Datalog Educational System column at all and it is therefore defined as a single fact without arguments This table can be used to compute arithmetics as e g DES SQL gt select 1 1 from dual answer a0 gt answer 2 Info 1 tuple computed As in MySQL DES also allows to omit the FROM clause in theses cases the compilation from SQL to Datalog adds the dual table as data source DES SQL gt select 1 1 answer a0 gt answer 2 Info 1 tuple computed Although this table is not displayed with the command dbschema it can be nevert
144. endency graph is useful for finding a stratification for the program ZCF 97 A stratification collects predicates into numbered strata NL A basic bottom up computation would solve all of the predicates in stratum 1 then 2 and so on until the meaning of the whole program is found With our approach we only resort to compute by stratum when a negative dependency occurs in the predicate dependency graph restricted to the query nevertheless each predicate that is actually needed is solved by means of the extension table mechanism described in the previous Fernando Sdenz P rez 165 204 Universidad Complutense de Madrid Datalog Educational System section As a consequence many computations are avoided w r t a naive bottom up implementation Outer join and aggregate goals are also collected into strata as if they were negative atoms in order to have their answer set completely defined and therefore ensure termination of the computation algorithm in presence of null values 5 16 4 Porting to Unsupported Systems DES is implemented with several Prolog files des pl des deg pl des_sql pl des_ra pl des_sql_debug pl des_dl_debug pl des_types pl des_tc pl and des_glue pl The first file contains the common predicates for all of the platforms both Prolog interpreters and operating systems following the Prolog ISO standard File des deg pl contains the definition of DCG expansion which varies from one system to another Files des
145. er view column_name type column_name type Fernando Sdenz P rez 156 204 Universidad Complutense de Madrid Datalog Educational System view column_name type column_name type view column_name type column_name type Seot Where view_name stands for view names column_name is a column name type is the column type and eot is the end of the transmission Remarks Return view schemas Views are returned alphabetically sorted Example Input tapi list_view_schemas Output v a number integer b string varchar 20 Seot e Command tapi list_table_constraints table_name Arguments table_name Table name enclosed between SQL delimiters if needed Answer NN PK CK CK FK FK FD FD IC IC Seot Where is a delimiter for different kinds of integrity constraints NN is a single line with the names of columns with existency constraint PK is a single line with the primary key constraint CK are candidate keys FK are foreign keys FD are functional dependencies IC are user defined integrity constraints and eot is the end of transmission Remarks Fernando Sdenz P rez 157 204 Universidad Complutense de Madrid Datalog Educational System List table constraints If there are no constraints of a given type no line is written Example Input tapi list_table_constraints s Output no existency constraint primary key b no
146. er SOULE Distribution ssanie ninn iaei i 14 EE EE 14 a2 Exec table Distrib ti n saitei ie bitens tanta autem de ae iien 14 22 22 Source DISHIDU NON stan cticncte ikon serran Aar ERa EE a Aai E 15 2 2 3 Starting DES from a Prolog interpreter sccchiscerccanceinireniniced Hoscacenscinines 15 3 Gettins Started ennen ae aae aea ETEO NES E S AE OET SS 15 Sech Dat log Mod sssroee raora E A E E E TE 16 32 SQL eet 19 3 3 Relational Algebra Meggie eegnen 22 EE TT 26 39 MCV CANS Bact acon ae ep teach clan TLE ang tak Soluce EE eae 27 3 6 Getinge Helper r eile n ea AA a ae aea eT Eaa Rae E A a AE TRES 27 4 Query Fottoen ee eege 27 AT D talog oriin en tees EAEE E EE AEE RE EEEE 28 ALM se Syntax mee a A A R 29 412 Rullesko i e T Ee 31 ANo E 31 e R BETE TEE 31 Tetris a RE e a AE ed T 32 4 1 6 Automatic Temporary Views erer 32 4 1 7 U d rscor d Variables sogen ienne Ee aa e EEn a een Ein 33 o EN e 110 TAE E 34 EN CM 36 4 1 10 UE 39 AV AT Outer JONS e E E E E EES 40 AVIZ SENSO TIES ALCS ee 42 4 1 12 1 Aggregate Functions ss ssssessssessesertsstssersstsesisntseenesesstnertestsnenestenentsnene 42 4 1 1222 Group by Predicates nn e aa dree 42 4 11 12 39 Aggreg te Predicates e re a e E E E E 45 4 113 SPU Chive EE 47 4 1 14 Integrity CoO straints E 48 AST VAL E 48 4 1 14 1 1 Types on Intensional Database is vcccccecisewrdivieniruicnscunnateesstdawcamannnet 50 4 1 14 1 2 Types on Propositional Relations 51 Fernand
147. ere SQLStatement is any SQL statement and LocalViewDefinitionl LocalViewDefinition1 are local view definitions that can only be used inside SQLStatement These local views are not stored in the database and are rather computed when executing SQLStatement Although they are local they must have different names from existing objects tables or views The syntax of a local view definition is as follows RECURSIVE ViewName Columnl ColumnN AS SQLStatement Here the keyword RECURSIVE for defining recursive views is not mandatory the parser simply ignores it Examples CREATE TABLE flights airline frm to departs arrives WITH RECURSIVE reaches frm to AS SELECT frm to FROM flights UNION SELECT r1 frm r2 to FROM reaches AS r1 reaches AS r2 WHERE r1 to r2 frm SELECT FROM reaches WITH Triples airline frm to AS SELECT airline frm to FROM flights RECURSIVE Reaches airline frm Col AS SELECT FROM Triples UNION SELECT Triples airline Triples frm Reaches to FROM Triples Reaches WHERE Triples to Reaches frm AND Triples airline Reaches airline SELECT frm to FROM Reaches WHERE airline UA EXCEPT SELECT frm to FROM Reaches WHERE airline AA 5 Adapted from GUW02 Fernando Sdenz P rez 71 204 Universidad Complutense de Madrid Datalog Educational System In addition shorter definitions for recursive views are allowed in DES The next view delivers the sa
148. ersidad Complutense de Madrid Datalog Educational System Info 2 tuples computed Next we enable the development mode for listings DES gt development on DES gt 1 X Y 1 1 1 1 2 SNULL 59 Info 2 tuples computed Here the internal representation of nulls is available If we request the listing of the stored rules in development mode DES gt listing p0 A NULL B P A not Sp1 A C p0 A B p1 A B p1 A B P A q B A B a C count p X X C 1 X Y p0 X Y p X X p X X q 1 e INI Info 8 rules listed Here we see several source to source transformations First the left join then the aggregate count and finally the disjunctive rule Development listings also allows to inspect the extension table looking at repeated facts involving nulls as follows DES gt assert q null DES gt assert q null DES gt q X q 1 q 3 q S NULL 64 q NULL 67 Fernando Sdenz P rez 123 204 Universidad Complutense de Madrid Datalog Educational System Info 4 tuples computed Compare this to the non development mode DES gt development off DES gt q X q 1 q 3 q null Info 3 tuples computed Also one can be aware from where nulls come because of their IDs as in DES gt assert p null DES gt listing p p NULL 70 P X xX p X X NIB Info
149. escendant gt such that ancestor is an ancestor of descendant the relation ancestor the set of tuples lt father child gt such that father is the father of child the relation father and the set of tuples lt mother child gt such that mother is the mother of child the relation mother tom grace No jack amy tony caroll woo au fred carolll carollII Figure 2 Family Tree The file family dl contains the following Datalog code which can be consulted with Ze family father tom amy father jack fred father tony carolII father fred carolIII mother grace amy mother amy fred Fernando S enz P rez 173 204 Universidad Complutense de Madrid Datalog Educational System mother carolI carolII mother carolII carolIII parent X Y father X Y parent X Y mother X Y ancestor X Y parent X Y ancestor X Y parent X Z ancestor Z Y The query ancestor tom X yields the following answer that is it computes the set of descendants of tom ancestor tom amy ancestor tom carolIII ancestor tom fred Info 3 tuples computed Solving the view son S F M father F S mother M S yields the following answer computing the set of sons Info Processing son S F M father F S mother M S son amy tom grace son carolII tony carolI son carollIII fred carolII son fred jack amy Info 4 tuples computed The
150. ey of Research on Deductive Database Systems Journal of Logic Programming 23 2 125 149 1995 Fernando S enz P rez 203 204 Universidad Complutense de Madrid Datalog Educational System SD91 Sae07 Shap83 SICStus Silv07 SRSS93 Tang99 TS86 Ullm95 VRK 91 Wiele WL04 ZCF 97 ZF97 C Shih and S W Dietrich Extension Table Evaluation of Datalog Programs with Negation Proceedings of the IEEE International Phoenix Conference on Computers and Communications Scottsdale AZ March 1991 pp 792 798 F Sdenz P rez ACIDE An Integrated Development Environment Configurable for LaTeX The PracTeX Journal 2007 Number 3 ISSN 1556 6994 August 2007 Shapiro E Algorithmic Program DeBugging ACM Distinguished Dissertation MIT Press 1983 SICS http www sics se sicstus Silva J A Comparative Study of Algorithmic Debugging Strategies in Proc of International Symposium on Logic based Program Synthesis and Transformation LOPSTR 2006 2007 pp 134 140 D Srivastava R Ramakrishnan S Sudarshan and P Seshadri Coral Adding Object Orientation to a Logic Database Language Proceedings of the International Conference on Very Large Databases 1993 Z Tang Datalog An Object Oriented Front End For The Xsb Deductive Database Management System http citeseer ist psu edu tang99datalog html H Tamaki and T Sato OLD
151. ez 187 204 Universidad Complutense de Madrid Datalog Educational System handling semistructured data in the context of Information Integration from the Semantic Web The NAIL project delivered a prototype with stratified negation well founded negation and modularity stratified negation Later it added the language Glue which is essentially single logical rules with SQL statements wrapped in an imperative conventional language PDR91 DMP93 The approach of combining two languages is similar to the aforementioned Coral which uses C It does not run on Windows platforms Another deductive database following this combination of declarative and imperative languages is Rock amp Roll BPFWD94 ADITI 2 VRK 91 is the last version of a deductive database system which uses the logic functional programming language Mercury It does not run on Windows platforms There is no further development planned for Aditi See also the Datalog entry in Wikipedia http en wikipedia org wiki Datalog 8 2 Technological Transfers Datalog has been extensively studied and is gaining a renowned interest thanks to their application to ontologies FHH04 semantic web CGL09 social networks RS09 policy languages BFGO7 and even for optimization GTZ05 Companies as LogicBlox Exeura Semmle and Lixto embody Datalog based deductive database technologies in the solutions they develop The high level expressivity of Datalog and its extensi
152. f that section if known or else a unique number Make the same adjustment to the section titles in the list of Invariant Sections in the license notice of the combined work In the combination you must combine any sections Entitled History in the various original documents forming one section Entitled History likewise combine any sections Entitled Acknowledgements and any sections Entitled Dedications You must delete all sections Entitled Endorsements 6 COLLECTIONS OF DOCUMENTS You may make a collection consisting of the Document and other documents released under this License and replace the individual copies of this License in the various documents with a single copy that is included in the collection provided that you follow the rules of this License for verbatim copying of each of the documents in all other respects You may extract a single document from such a collection and distribute it individually under this License provided you insert a copy of this License into the extracted document and follow this License in all other respects regarding verbatim copying of that document 7 AGGREGATION WITH INDEPENDENT WORKS A compilation of the Document or its derivatives with other separate and independent documents or works in or on a volume of a storage or distribution medium is called an aggregate if the copyright resulting from the compilation is not used to limit the legal rights of the compilation s users beyond
153. g error is found it is tried first as a SQL statement and second as an RA expression hypothetical Switch Enable or disable hypothetical queries on or off resp prolog Switch to Prolog interpreter all queries are parsed and executed in Prolog prolog Goal Trigger Prolog s SLD resolution for the goal Goal ra Switch to RA interpreter all queries are parsed and executed in RA ra Query Trigger RA evaluation for the query Query sql Switch to SQL interpreter all queries are parsed and executed in SQL sql SQL statement Trigger SQL resolution for SQL statement TAPI related See also Section 5 14 2 for more information tapi Input Process Input and format its output for TAPI communication Only a limited set of possible inputs are allowed cf Section 5 14 test_tapi Test the current TAPI connection TAPI enabled 0 Miscellanea check Switch Enable or disable integrity constraint checking on or off resp compact_listings Switch Enable or disable compact listings on or of resp Ferna ndo Sd enz P rez 147 204 Universidad Complutense de Madrid Datalog Educational System e display_answer Display whether display of computed tuples is enabled e display_answer Switch Enable or disable display of computed tuples on or off resp The number of tuples is still displayed e duplicates Switch Enable or disable integrity constraint checking on or off resp e negatio
154. gence delivered deductive databases Deductive database systems are database management systems built around a logical model of data and their query languages allow expressing logical queries Relational database languages where SQL is the de facto standard implement a limited form of logic whereas deductive database languages implement advanced forms of logic A deductive database is a system which includes procedures for defining deductive rules which can infer information in the so called intensional database in addition to the facts loaded in the so called extensional database The logic model for deductive databases is closely related to the relational model and in particular with the domain relational calculus Their query languages are related with the Prolog language and mainly with Datalog a Prolog subset without constructed terms in order to avoid infinite terms and other non declarative constructs such as the cut Origins of deductive databases can be found in automatic theorem proving and later in logic programming Minker Mink87 suggested that Green and Raphael GR68 were the pioneers in discovering the relation between theorem proving and deduction in databases They developed several question answer systems using a version of the Robinson resolution principle Robi65 showing that deduction can be systematically performed in a database environment Other pioneer systems were MRPPS MN82 DEDUCE 2 Chan78 and DADM KT
155. has been relaxed to LGPL version 3 The complete list of enhancements changes and fixed bugs are listed in Section 11 1 Fernando Sdenz P rez 8 204 Universidad Complutense de Madrid Datalog Educational System A novel contribution implemented in this system is a declarative debugger of Datalog queries CGS07 CGS08 which relies on program semantics rather than on the computation mechanism The debugging process is usually started when the user detects an unexpected answer to a query By asking questions about the intended semantics the debugger looks for incorrect program relations See Section 5 8 for details Also a similar declarative approach has been used to implement a SQL declarative debugger following CGS11b There possible erroneous objects correspond to views and the debugger looks for erroneous views asking the user whether the result of a given view is as expected In addition trusted views are supported to prune the number of questions This was extended to also include user information about wrong and missing tuples CGS12a See Section 5 9 for details In addition following the need for catching program errors when handling large amounts of data we also include a test case generator for SQL correlated views CGS10a Our tool can be used to generate positive negative and both positive negative test cases cf Section 5 10 1 1 Deductive Databases The intersection of databases logic and artificial intelli
156. heless dropped with a DROP TABLE SQL statement If it is deleted the just described behaviour is no longer possible In addition it cannot be redeclared with a CREATE TABLE SQL statement but with a type declaration as type dual Both DROP DATABASE statement and abolish command restore this table 4 2 6 2 Set SQL Queries The three set operators defined in the standard are available UNION EXCEPT and INTERSECT Also Oracle s MINUS is allowed as a synonymous for EXCEPT The first one also admits the form UNION ALL for retaining duplicates The syntax of a set SQL query is SQLStatement SetOperator SOLStatement Where SQLStatement is any SQL statement described in the data query part without any limitation SetOperator is any of the abovementioned set operators Examples SELECT FROM s UNION SELECT FROM t SELECT FROM s UNION ALL SELECT FROM t SELECT FROM s INTERSECT SELECT FROM t SELECT FROM s EXCEPT SELECT FROM t Note that parentheses are not mandatory in these cases and are only used for readability Fernando Sdenz P rez 70 204 Universidad Complutense de Madrid Datalog Educational System 4 2 6 3 WITH SOL Queries The WITH clause as introduced in the SQL 1999 standard and available in several RDBMS as DB2 Oracle and SQL Server is intended in particular to define recursive queries Its syntax is WITH LocalViewDefinitionl LocalViewDefinitionN SQLStatement Wh
157. ical value other terms are compared in Prolog standard order If a compound term is involved in a comparison operator it is evaluated as an arithmetic expression and its result is then compared for all operators by equality or unified for equality All comparison operators but equality demand ground arguments since they are not constraints but test operators and argument domains are infinite If a ground argument is demanded and a variable is received an exception is raised Fernando Sdenz P rez 87 204 Universidad Complutense de Madrid Datalog Educational System Next we list the available comparison operators where X and Y are terms variables constants or arithmetic expressions e X Y Syntactic equality Tests syntactic equality between X and Y It also performs unification when variables are involved This is the only comparison operator that does not demand ground arguments e X Y Syntactic disequality Tests syntactic disequality between X and Y e X gt Y Greater than Tests whether X is greater than Y e X gt Y Greater than or equal to Tests whether X is greater than or equal to than Y e xX lt Y Less than Tests whether X is less than Y e X lt Y Less than or equal to Tests whether X is less than or equal to Y 4 5 2 Datalog and Prolog Arithmetic Borrowed from most Prolog implementations arithmetic is allowed by using the infix operator is which is used to construc
158. ich should occur in Query Nulls are simply ignored e avg Query Variable Result Compute in Result the average of the numbers in the result set for the query Query and the attribute Variable which should occur in Query Nulls are simply ignored e min Query Variable Result Compute in Result the minimum of the numbers in the result set for the query Query and the attribute Variable which should occur in Query Nulls are simply ignored If there are no such numbers it returns null e max Query Variable Result Compute in Result the maximum of the numbers in the result set for the query Query and the attribute Variable which should occur in Query Nulls are simply ignored If there are no such numbers it returns nu11 4 5 8 Datalog Null related Predicates e is_null Term Succeed if Term is bound to a null value It raises an exception if Term is a variable e is_not_null Term Succeed if Term is not bound to a null value It raises an exception if Termis a variable 4 5 9 Duplicates The following built ins take effect when duplicates are enabled via the command duplicates on e distinct Query Succeed as many times as different ground answers are computed for Query e distinct Variables Query Succeed as many times as different ground tuples built with Variables are computed for Query Fernando S enz P rez 93 204 Universidad Complutense de Madrid Datalog Educational System 4 5 10 Top N Queries e t
159. ies and implements tabling mechanisms It runs both on Unix Linux and Windows operating systems Datalog Tang99 is a front end for the XSB deductive database system bddbddb WL04 stands for BDD Based Deductive DataBase It is an implementation of Datalog that represents the relations using binary decision diagrams BDDs BDDs are a data structure that can efficiently represent large relations and provide efficient set operations This allows bddbddb to efficiently represent and operate on extremely large relations IRIS Integrated Rule Inference System IRIS2008 is a Java implementation of an extensible reasoning engine for expressive rule based languages provided as an API Supports safe or un safe Datalog with locally stratified or well founded negation as failure function symbols and bottom up rule evaluation Coral RSSS94 is a deductive system with a declarative query language that supports general Horn clauses augmented with complex terms set grouping aggregation negation and relations with tuples that contain universally quantified variables It only runs under Unix platforms There is also a version which allows object oriented features called Coral SRSS93 FLORID F LOgic Reasoning In Databases KLW95 is a deductive object oriented database system supporting F Logic as data definition and query language With the increasing interest in semistructured data Florid has been extended for Fernando Sdenz P r
160. ined in the program family dl parent X Y father X Y mother X Y This clause is equivalent to parent X Y father X Y parent X Y mother X Y If you list the database contents via the command listing you will get the first form when development listings are off via the command development off Otherwise you get the second one command development on Datalog views and autoviews containing disjunctive bodies are allowed and the system informs about the program transformation needed to compute them For instance you can directly submit the rule above as a view at the DES prompt DES gt parent X Y father X Y mother X Y Info Processing parent X Y in the program context of the exploded query parent X Y father X Y parent X Y mother X Y parent amy fred parent carolI carolITI parent carolII carolIII parent fred carolIII parent grace amy parent jack fred Fernando Sdenz P rez 47 204 Universidad Complutense de Madrid Datalog Educational System parent tom amy parent tony carolIT Info 8 tuples computed 4 1 14 Integrity Constraints Integrity constraints allow to specify valid values for tuples in relations DES provides several predefined constraints stemmed from SQL type primary key and foreign key In addition a predefined functional integrity constraint is also provided Users can also define its own integrity
161. ing the first two DES Database commands for consulting and reconsulting files following Prolog syntax If a parameter is not accepted please try again enclosing it between single quotes 5 13 1 DES Database e FileNames Load the Datalog programs found in the comma separated list Filenames discarding both rules already loaded integrity constraints and SQL table and view definitions The extension table is cleared and the predicate dependency graph and strata are recomputed Examples Assuming we are on the examples distribution directory we can write DES gt mutrecursion family TAPI enabled See also consult Filename Fernando Sdenz P rez 138 204 Universidad Complutense de Madrid Datalog Educational System e FileNames Load the Datalog programs found in the comma separated list Filenames keeping rules already loaded integrity constraints and SQL table and view definitions The extension table is cleared and the predicate dependency graph and strata are recomputed TAPI enabled See also Filenames e abolish Delete the Datalog database This includes all the local rules including those which are the result of SQL compilations and external rules persisted predicates Integrity constraints and SQL table and view definitions are removed The extension table is cleared and the predicate dependency graph and strata are recomputed e abolish Name Delete the predicates matching Name This in
162. ion Example type d a string b int project b 1 d Duplicate elimination AR Return tuples in R discarding duplicates Concrete syntax distinct Relation Example distinct project a c Left outer join Ri aka Includes all tuples of Ri joined with matching tuples of Ro w r t condition d Those tuples of Ri which do not have matching tuples of R2 are also included in the result and columns corresponding to R2 are filled with null values Concrete syntax Relation1 ljoin Condition Relation2 Example a ljoin a b b Right outer join Ri ok Equivalent to R 9 Ri Concrete syntax Relation1 coin Condition Relation2 Fernando Sdenz P rez 84 204 Universidad Complutense de Madrid Datalog Educational System Example a rjoin a b b Full outer join Ri o R2 Equivalent to Ri R2xURi ok Concrete syntax Relation1 fjoin Condition Relation2 Example a fjoin a b b first each tuple in the group have the same values for attributes Gy G second matches condition possibly including aggregate functions and third is projected by expressions E En also possibly including aggregate functions Concrete syntax group_by GroupingAtts ProjectingExprs HavingCond Relation Example group_by a avg b min b gt 0 d 4 3 2 Recursion in RA Recursion in RA expressions can be specified by simply including the name of the view which is being defined in its defini
163. ions 2 built in comparison operators constants and variables Four examples of conditions are X gt 1 X Y X gt Y Y gt Z K lt Y Z lt 0 Note that X gt Z is now supported it can be solved whenever the rule where it occurs is safe cf Section 5 3 Relation functions A function has the form f al an where f is a function name ai are its arguments and maps to a relation Only built in functions are allowed The current provision of built in functions includes among others o not a Intended for computing the negation of its single argument a o 135 al1 a2 a3 Intended for computing the left outer join of the relations al left relation and a2 right relation committing the condition Boolean expression a3 join condition o xrj al a2 a3 Intended for computing the right outer join of the relations al left relation and a2 right relation committing the condition Boolean expression a3 join condition o 3 a1 a2 a3 Intended for computing the full outer join of the relations al left relation and a2 right relation committing the condition Boolean expression a3 join condition Note that outer join functions can be nested Literals Literals can be Fernando Sdenz P rez 30 204 Universidad Complutense de Madrid Datalog Educational System o Positive An atom o Negative A negated body of the form not Body where Body is a body cf next section Negative literals
164. ither Datalog queries nor SQL queries submitted from des are translated into external SQL and therefore processed by such EDB Only SQL queries in the same connection as the persisted predicate are processed by the EDB However future releases might translate queries submitted from des 5 2 9 6 Supported Platforms As stable versions of GNU Prolog and Ciao Prolog do not support ODBC connections persistency is not available in source distributions for these two systems Fernando S enz P rez 116 204 Universidad Complutense de Madrid Datalog Educational System 5 3 Safety and Computability 5 3 1 Classical Safety Built in predicates are appealing but they come at a cost which was already noticed in Section 4 5 The domain of their arguments is infinite in contrast to the finite domains of each argument of any user defined predicate Since it is neither reasonable nor possible to extensionally give an infinite answer when a subgoal involving a built in is going to be computed its arguments need to be range restricted i e the arguments have to take values provided by other subgoals To illustrate this point consider submitting the following view to the program file relop dl less X Y X lt Y c X Y Since the goal is less X Y and the computation is left to right both X and Y are not range restricted when computing the goal X lt Y and therefore this goal ranges over two infinite domains the one for X
165. lay the current predicate dependency graph restricted to the predicate with name PredName and Arity e pretty_ print Display whether pretty print listings is enabled e pretty_print Switch Enable or disable pretty print for listings on or off resp e referenced_relations Relation Display the name of relations that are directly referenced by a foreign key in relation Relation TAPI enabled e referenced_relations Relation Arity Display in format Name Arity those relations that are directly referenced by a foreign key in relation Relation Arity TAPI enabled Fernando Sdenz P rez 145 204 Universidad Complutense de Madrid Datalog Educational System e relation_exists relation_name Display t rue if the given relation exists and false otherwise TAPI enabled e relation_schema relation Dame Display relation schema of relation_name TAPI enabled e running_info Display whether running information as the incremental number of consulted rules as they are read is to be displayed e running_info Switch Enable or disable display of running information on or off resp e safe Display whether safety transformation is enabled e simplification Display whether program simplification is enabled e show_compilations Display whether compilations from SQL DQL statements to Datalog rules are to be displayed e show_compilations Switch Enable or disable display of extended information about compilation of SQ
166. length as with recursive path origin destination length as select edge 1 from edge union select path origin edge destination path length 1 from path edge where path destination edge origin and path length lt select count from edge select origin destination min length from path group by origin destination DES SQL gt select from spaths answer spaths origin spaths destination spaths length gt answer a a 2 answer a b 1 answer a c 1 answer a d 2 answer b a 1 answer b b 2 answer b c 2 answer b d 1 Info 8 tuples computed A possible RA formulation follows max_length max_length Fernando Sd enz P rez 172 204 Universidad Complutense de Madrid Datalog Educational System group_by count true edge path origin destination length project origin destination 1 edge union project path origin edge destination path length 1 path zjoin path destination edge origin and path length lt max_length edge product max_length spaths origin destination length group_by origin destination origin destination min length true path And its query ra select true spaths 6 4 Family Tree files family d1 sql ra This yet another classic program defines the family tree shown in Figure 2 the set of tuples lt parent child gt such that parent is a parent of child the relation parent the set of tuples lt ancestor d
167. light destination connect origin SELECT FROM connect answer connect origin string varchar connect destination string varchar gt answer lon ny answer mad ny answer mad par answer par ny Info 4 tuples computed One can use several assumptions in the same query but only one for a given relation If needed you can assume several rules by using UNION For example WITH flight origin destination time AS SELECT mad lon 2 0 UNION SELECT ny par 10 0 SELECT FROM travel which is equivalent to Fernando Sd enz P rez 74 204 Universidad Complutense de Madrid Datalog Educational System ASSUME SELECT mad lon 2 0 UNION SELECT ny par 10 0 IN flight origin destination time SELECT FROM travel Note SQL queries are only allowed as such i e they cannot be used as part of any view declaration Further versions might allow this 4 2 7 Information Schema Language ISL Several non standard statements are provided to display schema information e SHOW TABLES List table names TAPI enabled e SHOW VIEWS List view names TAPI enabled e SHOW DATABASES List database names TAPI enabled e DESCRIBE Relation Display schema for Relation as dbschema 4 2 8 SOL Grammar Here terminal symbols are Parentheses commas semicolons single dots asterisks and apostrophes Other terminal symbols are completely written in capitals as SELECT Percent
168. listed create or replace view CatsOrDogsOwner id aname specie as select O id P name P specie from Owner O Pet P PetOwner PO where O id PO id and P code PO code and specie cat or specie dog create or replace view CatsAndDogsOwner id aname as select A id A aname from CatsOrDogsOwner A CatsOrDogsOwner B where A id B id and A specie B specie create or replace view LessThan6 id as select id from CatsOrDogsOwner group by id having count lt 6 The intended answer of the views with the same name is kept In the case of CatOrDogOwner its intended answer is the multiset of owners with their pet names and species but limited to cats and dogs The very same computation tree as for pets1 sql results after replacing literals AnimalOwner by CatOrDogOwner However the new set of views is erroneous since the WHERE condition A specie B specie of CatsAndDogsOwner should beA specie lt gt B specie in order to ensure that the owner has at least one dog and one cat Now the user again detects an unexpected result from the view Guest since its outcome incorrectly includes the owner with identifier 4 Tom Cohen A new debugging session starts but now the old version of the views in the file pets_trust can be used as a trusted specification as follows DES gt process examples SQLDebugger pets2 sql DES gt debug_sql Guest trust_file examples SQLDebugger pets_trust Info Debugging view Guest 1 G
169. lt the example file relop dl DES gt cd examples DES gt consult relop dl Info 18 rules consulted where the default extension dl can be omitted Note that rules in files must end with a dot in contrast to command prompt inputs where the dot is optional in single line input Rules in a consulted file may span on multiple lines Then one can examine the contents of the database see Section 6 1 for an explanation of the consulted program via the command DES gt listing a al a a2 a a3 b al b b1 b b2 c al al c al b2 c a2 b2 cartesian X Y a X b Y difference X a X not b X full_join X Y fj a X b Y X Y inner_join X a X b X left_join X Y 1 See section 5 for more details about commands Fernando S enz P rez 17 204 Universidad Complutense de Madrid Datalog Educational System 15 a X b Y X Y projection X c X Y right_join X Y rj a X b Y X Y selection X a X X a2 union X a X b X Info 18 rules listed Submitting a query is pretty easy DES gt a X a al a a2 a a3 Info 3 tuples computed You can interactively add new rules with the command assert as in DES gt assert a a4 DES gt a X a al a a2 a a3 a a4 Info 4 tuples computed Saving the current database which may include such interactively added or deleted tuples is allow
170. ltiline Mode eege Geer dereen e EA ETE E A ies 122 Bo Developmett MOA gst iorri aaa e a e ine 122 57 Datalogand SQL Ee Ee ee 125 5 7 1 Tracing Datalg Queries Siristi innne sucutapepn cistucinictaaudaees eer 125 57 2 Tracing SQL Eeer EEN 126 5 8 Datalog Leed matgeet geren 127 5 9 SO Declarative Eeer 129 Fernando S enz P rez 5 204 Universidad Complutense de Madrid Datalog Educational System 5 9 1 Trusted Speci ca th OMS ee ee E 130 59 2 Missing anid Wrong TUplesiiiesco icsctemin terres croreraacrantioornvnecntnteeseo vieasnonies 132 SE eene eler 132 ig Wrong TUPleS eege 133 5 9 2 3 Displaying Extended Frbermarbten eege reese erter 134 5 10 SOL Test Case Generator cc itear aek re e a eek edi sees Gites 134 Et Batch PROCESS II eege 136 Dede APCS SAGES se sens E 137 5 19 Eege Seas cscs ens os eee sha cen nud nen ESE Roera EEEE eos eneg avast EEA 137 BSH RTR 138 5 13 2 ODBC Databases srini an eisni oteak envn rieien s esaka ESENE aiiin 141 5 13 3 Debugging and Test Case Generation sistas cieidiveeniin vecadilecearenundiion 141 513 4 Tabling porfis e e Eege 142 5 19 95 OOP RPA Syste EE 142 BNO Eege 143 5 13 7 ee EE 143 5 13 8 Query Languages ce antiquities EA 147 513 9 Ettel 147 5 13 10 eeh Eeer ener 147 5 13 11 Impl Mentor ien eien iren een e eE EE EE aE EAE ES SEEE EE SE aaea 148 5 14 T xt al A Pl csvecttasteianicncate mini ee 149 514 1 Notes about the Interface ee 150 5 T411 Ge 150 SIAT EE
171. ltiline on 3 1 Datalog Mode In this mode a query is sent to the Datalog processor If it does not follow Datalog syntax then it is sent first to the SQL processor see Section 4 2 and second to the RA processor see Section 4 3 should such query is written in any of these other query languages See caveats in Section 3 5 Commands see Section 5 13 are sent to the command processor Commands can end with an optional dot In single line mode Datalog inputs can also end with an optional dot but the dot is required in multi line Fernando Sdenz P rez 16 204 Universidad Complutense de Madrid Datalog Educational System mode Datalog mode is the default and can be anyway enabled via the command datalog The typical way of using the system is to write Datalog program files with default extension dl and consulting them before submitting queries Another alternative is to assert program rules from the system prompt Following the first alternative you write the program in a text file and then change to the path where the file is located by using the command cd Path where Path is the new directory relative or absolute Next the command consult FileName is used to consult the file FileName Provided there are a number or example files in the directory examples at the distribution directory and assuming that the current path is the distribution directory as by default one can use the following commands to consu
172. lutense de Madrid Datalog Educational System e license license A verbatim copy of the GNU Public License for this distribution 2 1 2 2 DES ACIDE Windows Bundle From the same URL above you can download a bundle including both DES and the integrated development environment ACIDE preconfigured to work with DES The following figure is a snapshot of the system File Edit Project View Confiquration Help EI mi Y A ec P kb ER consult process listing dbschema pdg strata abolish listet dear et cd Is pwd 5 9 DES3 0 HO cogregates w aggregates ra aggregates sq bom dl fact dl w family dl vw family ra vw family sql relop dl bom A Hl fomiya DI aggregates dl D aggregates ra relop dl D aggregates sq 2 Aggregates AE 2 Datalog Formulation EI relop dl s employee Name Department Salary 5 employee anderson accounting 1200 9 employee andrews accounting 1200 0 employee arlingon accounting 1000 EE EECH DES Datalog Educational System v 3 0 Type help for help about commands Fernando Saenz Perez c 2004 2012 GPD DISIA UCM Please send comments questions etc to fernan sip ucm es Web site http des sourceforge net This program comes with ABSOLUTELY NO WARRANTY is free software and you are welcome to redistribute it under certain conditions Type license for details ee te eee de d
173. ly after the copyright notices a license notice giving the public permission to use the Modified Version under the terms of this License in the form shown in the Addendum below G Preserve in that license notice the full lists of Invariant Sections and required Cover Texts given in the Document s license notice H Include an unaltered copy of this License I Preserve the section Entitled History Preserve its Title and add to it an item stating at least the title year new authors and publisher of the Modified Version as given on the Title Page If there is no section Entitled History in the Document create one stating the title year authors and publisher of the Document as given on its Title Page then add an item describing the Modified Version as stated in the previous sentence Fernando Sdenz P rez 196 204 Universidad Complutense de Madrid Datalog Educational System J Preserve the network location if any given in the Document for public access to a Transparent copy of the Document and likewise the network locations given in the Document for previous versions it was based on These may be placed in the History section You may omit a network location for a work that was published at least four years before the Document itself or if the original publisher of the version it refers to gives permission K For any section Entitled Acknowledgements or Dedications Preserve the Title of the section and p
174. m the table TableName all of its tuples matching the condition Condition It does not delete production rules asserted via assert Example DELETE FROM t WHERE a NOT IN SELECT a FROM s 4 2 6 Data Query Language There are three main types of SQL query statements SELECT statements set statements UNION INTERSECT and EXCEPT and WITH statements for building recursive queries Fernando Sdenz P rez 66 204 Universidad Complutense de Madrid 4 2 6 1 Basic SOL Queries Datalog Educational System The syntax of the basic SQL query statement is SELECT DISTINCT ALL ProjectionList FROM Relations WHERE Condition Where Square brackets indicate that the enclosed text is optional Also the vertical bar is used to denote alternatives ProjectionList is a list of comma separated columns or arithmetic expressions that will be returned as a tuple result Wildcards are allowed as for referring to all the columns in the data source and Relation for referring to all the columns in the relation Relation The name Relation can be the name of a table or an alias for a table or subquery Clause DISTINCT discards duplicates whereas clause ALL does not this is only noticeable when duplicates are enabled with the command duplicates on Condition is a logical condition built from comparison operators lt gt lt gt gt and lt Boolean operators AND OR and NOT Boolean constants TRUE FALSE
175. may raise a computing exception if non ground at run time Fernando Sdenz P rez 136 204 Universidad Complutense de Madrid Datalog Educational System Warning N2 is N 2 may raise a computing exception if non ground at run time Warning N1 is N 1 may raise a computing exception if non ground at run time Warning Next rule is unsafe because of variable s N fib N F N gt 1 N2 is N 2 fib N2 F2 N1 is N 1 fib N1 F1 F is F2 Fl DES gt f1ib 100 F fib 100 573147844013817084101 Info 1 tuple computed DES gt End log DES gt nolog 5 12 Messages DES system messages are prefixed by e Info An information message which requires no attention from the user Several information messages are hidden with the command verbose off which is the default mode e Warning A warning message which does not necessarily imply an error but the user is requested to focus on its origin These messages are always shown e Error An error message which requires attention from the user These messages are always shown e Exception An exception message which requires attention from the user These messages are always shown Examples of exception messages include instantiation errors and undefined predicates Prolog exceptions are caught by DES and shown to the user without any further processing Depending on the Prolog platform the system may continue by itself otherwise the user must type des incl
176. me result set as the first example above CREATE VIEW reaches frm to AS SELECT frm to FROM flights UNION SELECT r1 frm r2 to FROM reaches AS r1 reaches AS r2 WHERE rl1 to r2 frm 4 2 6 4 Hypothetical SOL Queries A novel addition to SQL in DES includes hypothetical queries Such queries are useful for instance in decision support systems as they allow to submit a query by assuming some knowledge which is not in the database Syntax of hypothetical queries is proposed as ASSUME LocalAssumptionI1 LocalAssumptionN SOLStatement Where SQLStatement is any SQL DQL statement and LocalAssumption1 LocalAssumptiomN are of the form DQLStatement IN ExistingRelation And LocalAssumptionN are added as unions to existing relations either tables or views Syntax of these local view definitions are as in WITH statements As an example let s consider a flight database defined by the following CREATE TABLE flight origin string destination string time real INSERT INTO flight VALUES lon ny 9 0 INSERT INTO flight VALUES mad par 1 5 INSERT INTO flight VALUES par ny 10 0 CREATE OR REPLACE VIEW travel origin destination time AS WITH connected origin destination time AS SELECT FROM flight UNION SELECT flight origin connected destination flight time connected time FROM flight connected WHERE flight destination connected origin SELECT FROM connected Here relation flight represen
177. mple is not stratified and in general we cannot ensure correctness for non stratifiable programs DES gt verbose on Info Verbose output is on DES gt c russell Info Consulting russell shaves barber M man M not shaves M M man barber man mayor shaved M shaves barber M end_of_file Info 4 rules consulted Info Computing predicate dependency graph Info Computing strata Warning Non stratifiable program 14 Remember that the system returns all of the possible solutions Fernando S enz P rez 179 204 Universidad Complutense de Madrid Datalog Educational System If we submit the query shaves X Y we get the positive facts as well as a set of undefined inferred information in our example whether the barber shaves himself as follows here verbose output is enabled DES gt shaves X Y Warning Unable to ensure correctness for this query shaves barber mayor Info 1 tuple computed Undefined shaves barber barber Info 1 tuple undefined If we look at the extension table contents by submitting the command list_et we get as answers Answers man barber man mayor not shaves mayor mayor shaves barber mayor Info 4 tuples in the answer set We can see that in particular we have proved additional negative information the mayor does not shaves himself and that no information is given for the undefined
178. mputed However if duplicates are enabled you get two answer tuples although the concrete values for the arguments of t are not visible DES gt duplicates on DES gt t _X _X Info Processing answer t _X _X answer answer Info 2 tuples computed 4 1 8 Negation DES ensures that negative information can be gathered from a program with negated goals provided that a restricted form of negation is used Stratified negation Ullm95 This broadly means that negation is not involved in a recursive computation path although it can use recursive rules The following program illustrates this point not b c d b Do D D The query a succeeds with the meaning a Observe also that not a does not succeed i e its meaning is the empty set DES provides two different algorithms for computing negation strata a default algorithm following a bottom up top down guided stratum saturation and et_not taken from SD91 which are selected via the command negation Algorithm cf Section 5 13 10 If you are interested in how programs with negation are solved for the algorithm strata you can find useful the following commands cf Section 5 13 7 DES gt pdg 3 In file negation dl1 located at the examples distribution directory Adapted from RSSWF97 Fernando Sdenz P rez 34 204 Universidad Complutense de Madrid Datalog Educational System Nodes d 0 a 0 b 0 c 0 Arcs a 0
179. n Algorithm Set the required Algorithm for solving negation strata or et not e halt Quit the system Synonyms quit q exit e e multiline Display whether multi line input is enabled e multiline Switch Enable or disable multi line input on or off resp e f output Switch Enable or disable display output on or off resp e process Filename Process the contents of Filename as if they were typed at the system prompt Extensions by default are sql and ini When looking for a file f the following filenames are checked in this order f f sql and E ini Synonyms p e safe Switch Enable or disable program transformation on or off resp e simplification Switch Enable or disable program simplification on or off resp Rules with equalities true and not BooleanValue are simplified e start_stopwatch Start stopwatch Precision depends on host Prolog system 1 second or milliseconds e stop_stopwatch Stop stopwatch Precision depends on host Prolog system 1 second or milliseconds e display_stopwatch Display stopwatch Precision depends on host Prolog system 1 second or milliseconds 5 13 11 Implementor e debug Enable debugging in the host Prolog interpreter Fernando Sdenz P rez 148 204 Universidad Complutense de Madrid Datalog Educational System e indexing Display whether hash indexing on extension table is enabled e indexing Switch Enable or disable hash
180. n ODBC Connection E 94 5 1 2 Heer Eege 95 51 3 Opening Several e EE 98 DLE WEE 99 5 1 5 Making a Connection EE 99 51 6 E EE 99 5 17 Schema and Data Meter EE 99 5 1 8 Integrity Constraints ODBC Connections and Dersietencn en 100 5 1 9 Caveats and imi ta E 102 GK veer 102 GE ODBC Metadata eegene eege EE EEN 103 GE ODBC Etranger 103 SEJA Platlorinespecine TISSUES sinini ib e 103 5 110 Vested OD BE LEE radiusen aore E EE wick 104 SNE E Le E 104 52 1 Persisting e EE 104 5 22 Using Persistent Predicates s an nie raer REE E a weedeat Ea 105 5 2 3 Processing a Persistency Assertion x csvssciceensesinntentneressearwecnatitennsveenseneaetionen sts 106 5 24 Restoring GOSS OM es darrii eeler ged 107 5 2 5 Schema of Persistent Predicates icitecesivensea eiert eet eer Sr 109 5 2 6 Removing Predicate Persisten Cy si cccuivacesnnsticcenccecmivka eege 110 52 7 Schema and Data Visi bit eege 111 9 28 ZAP PCa E 113 5 2 9 CAV CUS ee ee 115 92 9 1 Incomplete Meanings eer ege 115 5 2 9 2 Opening and Closing Connections eer eg 116 529 3 Eeer erregt 116 Ok ISU leede 116 5 2 9 5 External Database Processing civcecctes dideeceisncsasthandbepertaccnsofosetorapvecanciies 116 529 6 Supported Platf rMS errereen ger 116 5 37 Safetyand E EE eer 117 53 1 GEASS CAL Ale enee eege 117 5 3 2 Safety for Aggregates and Duplicate Elimination ccccceeeseseesesesees 120 5 4 Source to Source Transformations E 121 Did Mu
181. n order postorder or the default preorder test_case View Options Generate test case classes for the view View Options may include a class and or an action parameters The test case class is indicated by the values a11 positive negative the default positive or negative in the class parameter The action is indicated by the values display only display tuples the default replace replace contents of the involved tables by the computed test case or add add the computed test case to the contents of the involved tables in the action parameter tc_size Min Max Set the minimum and maximum number of tuples generated for a test case tc_size Display the minimum and maximum number of tuples generated for a test case tc_domain Min Max Set the domain of values for test cases between Min and Max tc_domain Display the domain of values for test cases Tabling clear_et Delete the contents of the extension table list_et List the contents of the extension table in lexicographical order First answers are displayed then calls list_et Name List the contents of the extension table matching Name First answers are displayed then calls list_et Name Arity List the contents of the extension table matching the pattern Name Arity First answers are displayed then calls Operating System cat Filename Type the contents of Filename enclosed between the following lines SS BEGIN AbsoluteFilename END Absolu
182. nd for input that the TAPI user must provide with a concrete input For example description for dropping a table includes tapi drop table table_name where table_name is the placeholder for your concrete table to be dropped e Lines starting with are remarks which are not needed to be included they are only for explanatory purposes e Types returned by a database or predicate handled by DES include o string varchar string varchar N string char N number integer number float DO Ome Where Nis an integer greater than 0 e Types returned by ODBC databases depend on the concrete external DBMS e Character strings as returned by DES are enclosed between single quotes This allows in particular to distinguish these strings from the null value which can occur in any data type e Datalog identifiers in TAPI inputs must be enclosed between single quotes should they contain special characters as blanks commas and quotes If an identifier contains a single quote this must be written twice as e g Peter ei which represents pete s e DDL Data Definition Language statements for SQL and Datalog include o CREATE TABLE SQL o CREATE VIEW SQL o RENAME SQL o strong constraint Datalog e DQL Data Query Language SQL statements include O SELECT o WITH e Any input to command tapi is processed as a DES input However output is only formatted for those commands and queries as listed in sections 5 14 2
183. nd or modify this document under the terms of the GNU Free Documentation License Version 1 3 or any later version published by the Free Software Foundation with no Invariant Sections no Front Cover Texts and no Back Cover Texts A copy of the license is included in the section entitled GNU Free Documentation License If you have Invariant Sections Front Cover Texts and Back Cover Texts replace the with Texts line with this with the Invariant Sections being LIST THEIR TITLES with the Front Cover Texts being LIST and with the Back Cover Texts being EEST If you have Invariant Sections without Cover Texts or some other combination of the three merge those two alternatives to suit the situation If your document contains nontrivial examples of program code we recommend releasing these examples in parallel under your choice of free software license such as the GNU General Public License to permit their use in free software Fernando S enz P rez 200 204 Universidad Complutense de Madrid Datalog Educational System Bibliography Agra88 A008 AOTWZ03 BCC97 BFG07 BPFWD94 Caba05 CGL09 CGS06b CGS07 CGS08 CGS10a R Agrawal Alpha An Extension of Relational Algebra to Express a Class of Recursive Queries IEEE Transactions on Software Engineering archive Volume 14 Issue 7 July 1988 P Ammann and J Offutt Introduction to Sof
184. ndition BWhereCondition UBWhereCondition HavingCondition As WhereCondition but including aggregate functions BWhereCondition WhereCondition UBWhereCondition TRUE Fernando Sdenz P rez 80 204 Universidad Complutense de Madrid Datalog Educational System ee See DQLstmt e WhereCondition Geetha aces ees NOT IN DQLstmt EE IS NOT NULL E NOT IN DQLstmt EE Operator ALL ANY WhereExpression WhereCondition AND OR WhereCondition WhereExpression Att Cte ArithmeticExpression DOLstmt AggrArithmeticExpression AttOrCte AVG MIN MAX SUM DISTINCT Att COUNT Att Att Cte Operator lt gt lt gt gt lt Cte Number String NULL Number is an integer or floating point number LEEESEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEES ISL Information Schema Language statements LESESEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEES ISLstmt SHOW TABLES Fernando Sdenz P rez 81 204 Universidad Complutense de Madrid Datalog Educational System SHOW VIEWS SHOW DATABASES DESCRIBE TableName ViewName 4 3 Extended Relational Algebra Following the seminal proposal Codd70 there have been some extensions to the basic and additional operators in the original proposal Here we include all the originals but division and extended operators for dealing with outer
185. new primary key constraint DES gt type q a int b int DES gt assert q 1 1 DES gt assert q 2 2 DES gt assert q 1 2 DES gt pk q a Error Primary key violation q a Offending values in database pk 1 Info Constraint has not been asserted 4 1 14 4 Candidate Key Uniqueness Constraint As a primary key a candidate key constraint specifies that no two tuples have the same values for a given set of columns Next a system session illustrates the use of a candidate key assertion DES gt type p a int b string DES gt ck p a Candidate key constraints are trivially satisfied when duplicates are disabled as relations are considered as sets irrespective of the current database instance that may contain duplicates for the arguments in the candidate key Several candidate key declarations are allowed for the same predicate and arity By contrast to primary keys several candidate key constraints are allowed for the same predicate DES gt ck p b DES gt ck p a b DES gt dbschema p Info Table p a number integer b string varchar NN a CK a CK b CK a b Fernando Sdenz P rez 52 204 Universidad Complutense de Madrid Datalog Educational System 4 1 14 5 Foreign Key A foreign key constraint specifies that the values in a given set of columns of a relation must exist already in the columns declared in the primary key constraint
186. nfo 2 tuples computed The execution of this goal allows to noting the basic differences between Prolog and Datalog engines First the former searches for solutions one by one that satisfy the goal projection X The latter gives the whole meaning of the user defined relation projection with the query projection X at a time And second note the default set oriented behaviour of the Datalog engine which discards duplicates in the answer 2 The meaning of a relation is the set of facts inferred both extensionally and intensionally from the program Fernando Sdenz P rez 26 204 Universidad Complutense de Madrid Datalog Educational System 3 5 Caveats Since the Datalog mode prompt accepts Datalog SQL and RA queries a given query can be interpreted in more than one language Let s consider the following system session in which a table is created and an RA query is submitted DES gt create table t a int DES gt distinct t Info Processing answer distinct t Warning Undefined predicate s t 0 Info 0 tuples computed Here we get an unexpected output coming from the Datalog interpreter as such input could be interpreted both as a Datalog query and an RA query To overcome such situations simply precede the query by the language selection command as follows DES gt ra distinct t answer t a number integer gt Info 0 tuples computed Alternatively switch to the other query
187. ng any large number of copies to give them a chance to provide you with an updated version of the Document 4 MODIFICATIONS You may copy and distribute a Modified Version of the Document under the conditions of sections 2 and 3 above provided that you release the Modified Version under precisely this License with the Modified Version filling the role of the Document thus licensing distribution and modification of the Modified Version to whoever possesses a copy of it In addition you must do these things in the Modified Version A Use in the Title Page and on the covers if any a title distinct from that of the Document and from those of previous versions which should if there were any be listed in the History section of the Document You may use the same title as a previous version if the original publisher of that version gives permission B List on the Title Page as authors one or more persons or entities responsible for authorship of the modifications in the Modified Version together with at least five of the principal authors of the Document all of its principal authors if it has fewer than five unless they release you from this requirement C State on the Title page the name of the publisher of the Modified Version as the publisher D Preserve all the copyright notices of the Document E Add an appropriate copyright notice for your modifications adjacent to the other copyright notices F Include immediate
188. ng if a rule is safe and if so it follows a program transformation for reordering its goals in order to make it computable in a left to right order This transformation does not come by default and it can be changed with the command safe Switch where Switch can take two values on for enabling program transformation and off for disabling this transformation If Switch is not included then the command informs whether program transformation is enabled or disabled The analysis performed by the system at compile time warns about safety and computability as follows 1 Raise an error if A goal involving a comparison operator will be non ground at run time b The expression E ina goal X is E will be non ground at run time c The goal not G contains unsafe variables or its safe variables are not restricted so far 2 Raise a warning if a A goal involving a comparison operator may be non ground at run time b The expression E ina goalX is E may be non ground at run time This analysis is performed in several cases e Whenever a rule is asserted either manually with the command assert or automatically when consulting programs A rule is always asserted even when it is detected as unsafe or it may raise an exception at run time Recall that safety is undecidable and there are rules detected as unsafe that can be actually and correctly computed e When a query conjunctive query autoview or view is submitted They are rejected
189. not computed from external data sources several features are not supported in the context of an opened ODBC connection e SQL tracer e Test case generator 5 1 9 4 Platform specific Issues ODBC connections are only supported by the provided binaries and the source distributions for SWI Prolog and SICStus Prolog The stable releases Ciao Prolog 1 14 2 and GNU Prolog 1 4 0 do not support this implementation If you use a 64 bit Windows OS notice that you can select to run either a 64 bit version of DES or a 32 bit one binaries built with SWI Prolog are provided in the download area In the first case 64 bit you must use the Database Connectivity ODBC Data Source Administrator tool Odbcad32 exe Fernando Sdenz P rez 103 204 Universidad Complutense de Madrid Datalog Educational System e The 32 bit version of the Odbcad32 exe file is located in the Yosystemdrive Windows SysWow6 4 folder e The 64 bit version of the Odbcad32 exe file is located in the Ysystemdrive Windows System32 folder Also notice that a 64 bit driver requires also a 64 bit database installation For instance you can define a 32 bit ODBC connection to 32 bit MS Access installation and a 64 bit ODBC connection to a 64 bit Oracle installation In this scenario both connectinos cannot be opened from the same DES instance which is either a 32 bit or 64 bit release 5 1 10 Tested ODBC Drivers Several data sources have been successfully tested on
190. nother example is DES gt X ise 2 718281828459045 is exp 1 Info 1 tuple computed DES gt e is e Info 0 tuples computed This means that the built in arithmetic constant e cannot be used outside of an arithmetic expression and it is otherwise understood as a user defined relation Here an input error is not raised since e could be a user defined relation In fact this should raise a type error but they are not currently controlled In addition note that arithmetic expressions are compound terms which are translated into an internal equivalent representation The last example shows this since the constant e is translated to exp 1 Concluding the infix infinite relation is is understood as the set of pairs lt V E gt such that V is the equivalent value to the evaluation of the arithmetical expression E Note that since this relation is infinite we may reach non termination Let s consider the following program loop d1 in the distribution directory with the query Loop X loop 0 loop X loop Y X is Y 1 Evaluating that query results in a non terminating cycle because unlimited tuples is N N 1 become computed To show it try the query press Ctrl C and type listing et at the Prolog prompt only when DES has been started from a Prolog interpreter 4 5 3 SQL Arithmetic Arithmetic expressions are constructed with the arithmetic operators listed in the next section They ar
191. ns stated herein The Document below refers to any such manual or work Any member of the public is a licensee and is addressed as you You accept the license if you copy modify or distribute the work in a way requiring permission under copyright law A Modified Version of the Document means any work containing the Document or a portion of it either copied verbatim or with modifications and or translated into another language A Secondary Section is a named appendix or a front matter section of the Document that deals exclusively with the relationship of the publishers or authors of the Document to the Document s overall subject or to related matters and contains nothing that could fall directly within that overall subject Thus if the Document is in part a textbook of mathematics a Secondary Section may not explain any mathematics The relationship could be a matter of historical connection with the subject or with related matters or of legal commercial philosophical ethical or political position regarding them The Invariant Sections are certain Secondary Sections whose titles are designated as being those of Invariant Sections in the notice that says that the Document is released under this License If a section does not fit the above definition of Secondary then it is not allowed to be designated as Invariant The Document may contain zero Invariant Sections If the Document does not identify any Invariant Sections then
192. nt state of the implementation the following conditions must hold for a rule to be made persistent The rule does not contain calls to built ins but comparison operators The rule does not form a recursive cycle Nonetheless the rule is kept in the in memory database for computing the meaning of the predicate when needed This is performed by the deductive engine which couples the processing of the external database with its own processing to derive the meaning of the predicate Therefore all the deductive computing power is preserved although the external persistent media lacks some features as for instance recursion think of MySQL and MS Access Anyway such rules which are not projected to the external database are stored on it as metadata information This is needed to restore the complete definition of a persistent predicate upon restoring c f next section Further releases might contain relaxed conditions Any time a predicate is made persistent its associated connection is opened if it not was opened already the current connection is not changed anyway The connection is not closed even when you drop the assertion see Section 5 2 6 5 2 4 Restoring a Session As expected if you make a predicate persistent and quit DES in a next session you can recover the state of this predicate It is simply done by submitting again the same assertion as used to make the predicate persist for the first time However note that any
193. nt to the command processor SQL queries can end with an optional semicolon in single line mode Multi line mode requires the ending semicolon SQL mode is enabled via the command sql Datalog and RA queries cannot be handled by this mode If we want to develop an analogous SQL example session to the Datalog example in the last section we can submit the first inputs also available in the file examples relop sql1 listed below the example is augmented to provide a first glance of SQL Now answer relations to SQL queries are denoted by the relation name answer Also note that lines starting by are simply remarks If you wish to automatically reproduce the following interactive session of inputs you can type process examples relop sql notice that you must omit examples if you are in this directory already Info Processing file relop sql DES gt Switch to SQL interpreter DES gt sql DES SQL gt Creating tables DES SQL gt create or replace table a a string DES SQL gt create or replace table b b string DES SQL gt create or replace table c a string b string DES SQL gt Listing the database schema DES SQL gt dbschema Info Table s a a string varchar Fernando Sdenz P rez 19 204 Universidad Complutense de Madrid Datalog Educational System b b string varchar c a string varchar b string varchar Info No views Info No integrity constraints DES SQL gt Inserting values into
194. o Constraint successfully parsed Info Computing predicate dependency graph Info Computing strata Info Checking user defined integrity constraint over database count edge A B Es count node N Ns D is Ns Es D 1 Info Computing by stratum of edge A B node A Info Computing predicate dependency graph Info Computing strata DES gt assert edge e f An unconnected component Info Checking user defined integrity constraint over database count edge A B Es count node N Ns D is Ns Es D 1 Info Computing by stratum of edge A B node A Info Computing predicate dependency graph Info Computing strata Error Integrity constraint violation ic Es Ns D count edge A B Es count node N Ns D is Ns Es D 1 Offending values in database ic 4 6 2 User defined integrity constraints are dropped when abolishing the database or consulting a file 4 1 14 8 Dropping Constraints Any predefined or user defined integrity constraint can be dropped with the command drop_ic see Section 5 13 1 followed by the constraint to be dropped with the same syntax as its declaration Fernando S enz P rez 58 204 Universidad Complutense de Madrid Datalog Educational System 4 1 14 9 Caveats Either by consulting a program or by dropping the current database or by abolishing the database all integrity constraints are removed including
195. o S enz P rez 3 204 Universidad Complutense de Madrid Datalog Educational System 4 1 14 2 Nullability Existency Constraint s cc cscesstessessvaveondeessncesssectess Seu 51 AF143 EE 51 4 1 14 4 Candidate Key Uniqueness Constant 52 4 1145 Foreisn KEYone aee ee ee E AE ARER 53 4 1 14 6 Functional Dependency iss stzccavcesssievcorensbeasnadiiysiaenouecanciaiteptienivcdtelineathenies 54 4 1 14 7 User defined Integrity Constraints sccccsce casissteroieca cichesteseorees dhostartianies 55 4 1 14 8 Dropping Constraints hsna a a 58 4 1 14 97 Caveits Ee EE EE 59 E 59 42 1 Main E e 59 42 2 Main RATT OS ege EE e Eri EE Reis ta skeen A rE EE ven et 60 ADD E EE 61 4 2 4 bereede 61 424 Creating TA EE 61 AD Ad Eet MER dee 63 424 3 Dropping E 64 42 44 Dropping Mie Ee 65 42 4 Renaming VAL GS sc eene Eeer 65 4 2 4 6 Renaming VIC WS a sot dos connec seats tenant Ua dutas teisesse iniret iriaren 65 ADA Dropping Databases snenia ereinen en ein ioiei 65 4 2 5 Data Manipulation Late 65 4 25 11 RE iniia a a tesla sapiens 65 25 22 Deleting dE eet 66 42 6 Data Query Etgen Eed 66 42 6 Basic SCHEI elt terest have rats esd ee ee ee a i 67 EN WE ET 69 420 1 2 EA PADI G cacy teadicsicresndesee des nea ae eaeoe Mes R A EN eaae 69 4 2 6 Get SQL Queries S neea eech ee dee Seed dee See dds 70 426 3 WITH SOL Ou TS eee nite a ee es 71 22 Hypothetical EE 72 4 2 7 Information Schema Language ISL ciscsccvcnctsccsstecovcniecs
196. o have the same name that the column of s where the constraint applies Ge b So an error is thrown because columns s b and t b have different types DES SQL gt CREATE OR REPLACE TABLE s a INT b INT REFERENCES t PRIMARY KEY a b Error Type mismatch s b number int lt gt t b string varchar Error Imposing constraints A declared primary key or foreign key constraint is checked whenever a new tuple is added to a table following relational databases Note that assertion of rules from the Datalog side are allowed but not checked A Datalog rule should be viewed as a component of the intensional database RDBs avoid to define a view with the same name as a table and therefore there is no way of unexpected behaviours such as the illustrated below DES SQL gt create or replace table t a int b int c int d int primary key a c DES SQL gt insert into t values 1 2 3 4 Info 1 tuple inserted DES SQL gt The following is expected to raise an error DES SQL gt insert into t values 1 1 3 4 Error Primary key violation when trying to insert t 1 1 3 4 Info 0 tuples inserted DES SQL gt However the following is allowed Fernando Sdenz P rez 62 204 Universidad Complutense de Madrid Datalog Educational System DES SQL gt assert t X Y Z U X 1 Y 2 2 3 U 4 DES SQL gt listing t 1 2 3 4 t X Y Z U X 1 A GNK AWN Production rules those defining the intensional
197. o the call table whenever they are solved This allows us to detect whether a call has been previously solved and we can use the results in the extension table if any The algorithm which implements this idea can be sketched as follows First test whether there is a previous call that subsumes the current call There are two possibilities 1 there is such a previous call then use the result in the answer table if any It is possible that there is no such a result for instance when computing the goal p in the program p p and we cannot derive any information 2 otherwise process the new call knowing that there is no call or answer to this call in the extension table So firstly store the current call and then solve the goal with the program rules recursively applying this algorithm Once the goal has been solved if succeeded store the computed answer if there is no any previous answer subsuming the current one note that through recursion we can deliver new answers for the same call This so called memoization process is implemented with the predicate memo 1 in the file des p1 of the distribution and will also be referred to as a memo function in the rest of this manual Negative facts are produced when a negative goal is proved by means of negation as failure closed world assumption In this situation a goal as not p which succeeds produces the fact not p which is added to the answer table just the same as proving a po
198. of a table t can take values between two integers one can use the SQL clause CHECK in the creation of the table as follows CREATE TABLE t c INT CHECK c BETWEEN O AND 10 In contrast in Datalog you can submit the following constraints DES gt type t c int DES gt t X X lt 0 X gt 10 Notice that the rule body succeeds for values in t out of the interval 0 10 So an integrity constraint specifies unfeasible values rather than feasible Also note that whilst several predefined constraints are allowed in a constraint only one user defined integrity constraint is allowed A couple of assertions to show the behaviour of the above example follow DES gt assert t 0 DES gt assert t 11 Error Integrity constraint violation ic X t X x lt 0 X gt 10 Offending values in database ic 11 Note that to be able to interpret that offending values the integrity constraint is shown as a rule defining a new predicate ic where the rule s head has as many variables as relevant variables in the constraint Then offending values are encapsulated in the meaning of the constraint relation ic A rule body of a constraint is any valid rule body i e goals in constrainsts can refer to other user defined or built in predicates as well including negation 4 This CHECK SQL clause is not yet supported by DES Fernando Sdenz P rez 55 204 Universidad Complutense de Madrid Datalog Educational System
199. of freedom to assure everyone the effective freedom to copy and redistribute it with or without modifying it either commercially or noncommercially Secondarily this License preserves for the author and publisher a way to get credit for their work while not being considered responsible for modifications made by others This License is a kind of copyleft which means that derivative works of the document must themselves be free in the same sense It complements the GNU General Public License which is a copyleft license designed for free software We have designed this License in order to use it for manuals for free software because free software needs free documentation a free program should come with manuals providing the same freedoms that the software does But this License is not limited to software manuals it can be used for any textual work regardless of subject matter or whether it is published as a printed book We recommend this License principally for works whose purpose is instruction or reference Fernando Sdenz P rez 193 204 Universidad Complutense de Madrid Datalog Educational System 1 APPLICABILITY AND DEFINITIONS This License applies to any manual or other work in any medium that contains a notice placed by the copyright holder saying it can be distributed under the terms of this License Such a notice grants a world wide royalty free license unlimited in duration to use that work under the conditio
200. of the form Anna 1 1 y n a Iyl y Info Debugging view standard Input Should standard include a tuple of the form Anna 2 1 y n a yl y Info Debugging view standard Input Should standard include a tuple of the form Anna 3 0 y n a yl y Info Buggy view found intensive The first answer m Anna indicates that Anna is missing in the view awards Next the user indicates that view intensive should not include Anna The debugger then asks three simple questions involving the view standard After checking the information for Anna the user indicates that the listed tuples are correct Then the tool points out intensive as the buggy view after only five simple questions Observe that intermediate views can contain hundreds of thousands of tuples but the slicing mechanism helps to focus only on the source of the error 5 9 2 2 Wrong Tuples Let s consider a modification of the database defined in awards1 sql as found in file awards2 sql where the view basicLevelStudents has been incorrectly defined We process this file inspect the outcome of awards and notice that Anna should not be in the result set Then we proceed with the debugging session as follows DES gt process examples SQLDebugger awards2 DES gt debug_sql awards Info Debugging view awards Fernando Sdenz P rez 133 204 Universidad Complutense de Madrid Datalog Educational System 1 awards Ana
201. oined as in the next equivalent query DES gt 1j a X b Y true X Y Info Processing answer X Y 1j a X b Y true X Y answer al al Info 1 tuple computed Outer join relations can be nested as well DES gt 13 a X rj b Y c U V Y U X Y Info Processing Fernando Sdenz P rez 41 204 Universidad Complutense de Madrid Datalog Educational System answer X Y U V 15 a X rj b Y c U V U X Y answer al al al ali answer al al al D i answer a2 null null null answer a3 null null null Info 4 tuples computed Note that compound conditions must be enclosed between parentheses as in DES gt 135 a X c U V X gt U X gt V Info Processing answer X U V in the program context of the exploded query answer X U V 1j a X c U V X gt U X gt V answer al null null answer a2 al al answer ai al b i answer a3 al al answer ai al b i answer a3 a2 b2 Info 6 tuples computed 4 1 12 Aggregates Aggregates refer to functions and predicates that compute values with respect to a collection of values instead of a single value Aggregates are provided by means of five usual computations sum cumulative sum count element count avg average min minimum element and max maximum element In addition the less usual times cumulative product is also provided They behave close to most SQL implementations
202. olog executable path depending on the Prolog interpreter you use a Ciao Prolog 1 ciaorc b GNU Prolog entry goal des p1 c SlCStus Prolog 1 des pl d SWI Prolog g ensure_loaded des remove win_app if present Another alternative is to write a batch file similar to the script file described in the next section 2 2 2 Linux 2 2 2 1 Executable Distribution You can create a script or an alias for executing the file des at the distribution root This executable has been generated under SICStus Prolog so that all SlCStus notes in the rest of this document also apply to these executables In addition since it is a portable application it needs to be started from its distribution directory Fernando Sdenz P rez 14 204 Universidad Complutense de Madrid Datalog Educational System 2 2 2 2 Source Distribution You can write a script for starting DES according to the selected Prolog interpreter as follows a Ciao Prolog SCIAO 1 ciaorc Provided that CIAO is the variable which holds the absolute filename of the Ciao Prolog executable b GNU Prolog GNU entry goal des pl Provided that GNU is the variable which holds the absolute filename of the GNU Prolog executable c SICStus Prolog SICSTUS 1 des pl Provided that SICSTUS is the variable which holds the absolute filename of the SICStus Prolog executable d SWI Prolog SWI g ensure_loaded des P
203. ommands temporary views and conjunctive queries see next sections If an error leads to an exit from DES and you have started from a Prolog interpreter then you can write des without the double quotes and with the dot at the Prolog prompt to continue Although a query in any of the languages above can be submitted from such prompt there are currently four modes available which enable to use a concrete query interpreter for Datalog SQL Relational Algebra and Prolog The first one is the default A mode can be switched via the commands datalog sql ra and prolog respectively Note that commands always start with a slash Anyway if you are ina given mode you can submit queries or goals to other interpreters simply writing the query or goal after any of the previous commands Also if you are in Datalog mode you can directly submit both SQL and RA queries Data are stored in a deductive database including facts and rules All queries and goals irrespective of the language refer to this database When an external database is opened see Section 5 1 their tables and views are available and can be queried from Datalog Prolog and SQL In contrast with other interpreters default input mode is single line which means that the input will be processed after hitting the Intro key which allows to omit the terminating character Nonetheless this mode can be switched to multi line as described in Section 5 5 with the command mu
204. on of the first argument of c projection X c X Y sigma X a2 a Selecting tuples from a such that its first argument is a2 selection X a X X a2 axb Cartesian product of relations a and b cartesian X Y a X b Y a x b Natural inner join of relations a and b inner_join X a X b X a x b Left outer join of relations a and b left_join X Y 1j a X b Y X Y a x b Right outer join of relations a and b right_join X Y rj a X b Y X Y a x b Full outer join of relations a and b full_join X Y fj a X b Y X Y S aUb Set union of relations a and b union X a X b X a b Set difference of relations a and b difference X a X not b X Once the program is consulted you can query it by for example DES gt projection X projection al projection a2 Info 2 tuples computed The result of a query is the meaning of the view i e the fact set for the query derived from the program whether intensionally or extensionally In the above example projection X corresponds to the projection of the first argument of relation c Fernando Sdenz P rez 167 204 Universidad Complutense de Madrid Datalog Educational System The second view in Section 4 1 5 returns Info Processing a X b X a al a a2 a a3 a bl a b2 Info 5 tuples computed For abolishing this program and execute th
205. onnections and Persistency Integrity constraints as described in Section 4 1 14 are monitored by DES for the local deductive database This means that inserting values directly into external tables either by submitting an INSERT INTO statement from the opened connection or by inserting values out of DES is not monitored for constraint consistency However as constraint consistency checking considers all visible data when asserting into the local database data from the current opened connection is also taken into account The following system session shows a possible scenario illustrating these situations DES gt Zuse db des DES gt create or replace table t a int primary key DES gt dbschema Info Database Sdes Info Table s t a number integer PK a Info No views Fernando Sdenz P rez 100 204 Universidad Complutense de Madrid Datalog Educational System Info No integrity constraints DES gt open_db mysql Table t is also an external table in connection mysql DES gt dbschema t Info Database mysql Info Table t a integer 4 Retrieve tuples from external table t DES gt select from t answer a integer 4 gt Info 0 tuples computed The following is inserted in external table t Recall that SQL statements under an opened connection are submitted directly to the external RDBMS DES gt insert into t values 1 Info 1 tuple inserted DES gt inse
206. ons This section lists some caveats and limitations of the current implementation of ODBC connections to external data sources 5 1 9 1 Caching Data in relational tables are cached in the memo table during Datalog computations and it is not requested anymore until this cache is cleared either explicitly with the command clear_et or because a command or statement invalidating its contents as a SQL update query Therefore it could be possible to access outdated data from a Datalog query Let s consider DES SQL gt datalog t X t 1 Info 1 tuple computed Then from the MySQL client mysql gt insert into t values 2 Query OK 1 row affected 0 06 sec And after in DES the new tuple is not listed via a Datalog query DES SQL gt datalog t X t 1 Info 1 tuple computed However a SQL statement returns the correct answer DES SQL gt select from t answer a varchar gt answer 1 answer 2 Info 2 tuples computed In addition it is not recommended to mix Datalog and SQL data It is possible to assert tuples with the same name and arity as existing RDBMS s tables and or views Let s consider the same table t as above with the same data two tuples t 1 and t 2 and assert a tuple t 3 as follows DES SQL gt assert t 3 DES SQL gt datalog t X Fernando Sdenz P rez 102 204 Universidad Complutense de Madrid Datalog Educational System
207. ons has therefore been acknowledged as a powerful feature to deal with knowledge based information The first commercial oriented deductive database system was the Smart Data System SDS and its declarative query language Declarative Reasoning DECLARE KSSD94 with support for stratified negation and sets Currently XSB and DLV have been projected to spin off companies and they develop deductive solutions to real world problems 9 Future Enhancements The following list in order of importance suggests some points to address for enhancing DES e Multiple DB connections e Disjunctive heads e Information about cycles involving negation in the loaded program e Complete algorithm for finding undefined information e Constraints reals integers enumerated types e Precise error reporting for SQL and Datalog syntax errors If you find worthwhile for your application either some of the points above or others not listed please inform the author for trying to guide the implementation to the most demanded points Fernando Sdenz P rez 188 204 Universidad Complutense de Madrid Datalog Educational System 10 Caveats and Limitations Datalog o No compound terms as arguments in user relations o Termination is ensured up to arithmetic There is no provision for numerical bounds o No database updates via Datalog rules are allowed o Rules in consulted files must end with a dot in contrast to command prompt inputs in singl
208. op N Query Succeed at most N times for Query 5 System Description This section includes descriptions about the connection to relational database systems via ODBC connections persistency safety and computability issues source to source transformations the declarative debuggers and tracers the batch processing system messages and finally lists all the available commands 5 1 RDBMS connections via ODBC DES provides support for connections to relational database management systems RDBMSs in order to provide data sources for relations This means that a relation defined in a RDBMS as a view or table is allowed as any other relation defined via a predicate in the deductive database Then computing a query can involve computations both in the deductive inference engine and in the external RDBMS SQL engine Such relations become first class citizens in the deductive database and therefore can be queried in Datalog and RA If the relation is a view it will be processed by the SQL engine When an ODBC connection is opened all SQL statements are redirected to such connection so DES does not longer process such statements This means that all the SOL features of the connected RDBMS are available Almost any relational database RDB can be accessed from DES using an ODBC connection Relational database management system RDBMS manufacturers provide ODBC implementations which run on many operating systems Microsoft Windows Linux Mac
209. over Text and one of Back Cover Text may be added by or through arrangements made by any one entity If the Document already includes a cover text for the same cover previously added by you or by arrangement made by the same entity you are acting on behalf of you may not add another but you may replace the old one on explicit permission from the previous publisher that added the old one The author s and publisher s of the Document do not by this License give permission to use their names for publicity for or to assert or imply endorsement of any Modified Version 5 COMBINING DOCUMENTS You may combine the Document with other documents released under this License under the terms defined in section 4 above for modified versions provided that you include in the combination all of the Invariant Sections of all of the original documents unmodified and list them all as Invariant Sections of your combined work in its license notice and that you preserve all their Warranty Disclaimers Fernando Sdenz P rez 197 204 Universidad Complutense de Madrid Datalog Educational System The combined work need only contain one copy of this License and multiple identical Invariant Sections may be replaced with a single copy If there are multiple Invariant Sections with the same name but different contents make the title of each such section unique by adding at the end of it in parentheses the name of the original author or publisher o
210. processed Duplicates are disabled by default i e answers are set oriented But they can be enabled as well which is useful in Datalog SQL and RA queries see Section 4 1 9 For instance DES Prolog gt duplicates on Info Duplicates are on DES Prolog gt datalog projection X projection al projection al projection a2 Info 3 tuples computed 3 3 Relational Algebra Mode In this mode queries are sent to the Relational Algebra RA processor whereas commands cf Section 5 13 are sent to the command processor RA queries can end with an optional semicolon in single line mode Multi line mode requires the ending semicolon RA mode is enabled via the command ra Datalog and SQL queries cannot be handled by this mode If we want to develop an analogous RA example session to the former examples we can submit the first inputs also available in the file Fernando S enz P rez 22 204 Universidad Complutense de Madrid Datalog Educational System examples relop ra listed below Now answer relations to RA queries are denoted by the relation name answer As before lines starting by either or are simply remarks If you wish to automatically reproduce the following interactive session of inputs you can type process examples relop ra notice that you must omit examples if you are in this directory already DES RA gt Testing the just inserted values DES RA gt select true a answer a a s
211. rative debugger e des_types pl Type inferrer for SQL RA and Datalog e des_tc pl Test case generator for SQL views e des_glue pl Contains particular code for the selected host Prolog system e ciaorc Only for Ciao Prolog system Contains initialization code for this system e doc manualDES lt version gt pdf This manual e examples dl Example files which will be discussed in Section 6 e license license A verbatim copy of the GNU Public License for this distribution 2 1 2 Executable Distribution 2 1 2 1 Windows From the same URL above you can download a Windows executable distribution in a single archive file containing the following e readmeDES lt version gt txt A quick installation guide and file release contents e des exe Console executable file intended to be started from a OS command shell as depicted in the next figure Fernando Sdenz P rez 10 204 Universidad Complutense de Madrid Datalog Educational System B C des des exe cE 2s gt EE IE HE DE JE JE DE JE JE JE DE DE JE JE DE JE JE DE DE JE JE JE JE DE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE JE J FE F DES Datalog Educational System v 3 0 Type help for help about commands Fernando Saenz Perez c 2004 2012 GPD DISIA UCM Please send comments questions etc to fernan sip ucm es Web site http des sourceforge net KK KK KK K K K K XK x x This program comes with ABSOLUT
212. re computed To illustrate this consider the query b for the same program DES computes the predicate dependency subgraph for b i e all of the predicates which are reachable from b and then a stratification is computed Notice the different information given by the system for solving the queries a and b here verbose output is currently enabled with the command verbose on DES gt a Info Computing by stratum of b a Info 1 tuple computed DES gt b Info 0 tuples computed Fernando Sdenz P rez 35 204 Universidad Complutense de Madrid Datalog Educational System For the goal a the system informs that b is previously computed nevertheless taking advantage of the extension table mechanism whereas for the goal b there is no need of resorting to the stratum by stratum solving Finally consult also Section 5 3 for limitations in the use of negation 4 1 9 Duplicates Duplicates in answers are removed by default However it is also possible to enable them with the command duplicates on This allows to generate answers as multisets instead of as the typical set oriented deductive systems behave Computing the meaning of a relation containing duplicates in the extensional database i e its facts will include all of them in the answer as in DES gt duplicates on DES gt assert t 1 DES gt assert t 1 DES gt t X t 1 t 1 Info 2 tuples computed Rules can also be source of
213. red carolIII father tony carolIT Info 4 tuples in the answer table Info Remaining views mother 2 Input Continue y n yl Info Tracing view mother mother amy fred mother grace amy Info 4 tuples in the answer table Info No more views to trace DES SQL gt trace_datalog father X Y Info Tracing predicate father father fred carolIII father tony carolIT Info 4 tuples in the answer table Info No more predicates to trace 5 8 Datalog Declarative Debugger Our approach CGS07 to debug Datalog programs is anchored to the semantic level instead of the computation level We have implemented a novel way of applying declarative debugging also called algorithmic debugging a term first coined in the logic programming field by E H Shapiro Shap83 to Datalog programs With this approach it is possible to debug queries and diagnose missing answers an expected tuple is not computed as well as wrong answers a given computed tuple should not be computed Our system uses a question answering procedure which starts when the user detects an unexpected answer for some query Then if possible it points to the program fragment responsible of the incorrectness Fernando Sdenz P rez 127 204 Universidad Complutense de Madrid Datalog Educational System The debugging process consists of two phases During the first phase the debugger builds a computation graph CG for th
214. reserve in the section all the substance and tone of each of the contributor acknowledgements and or dedications given therein L Preserve all the Invariant Sections of the Document unaltered in their text and in their titles Section numbers or the equivalent are not considered part of the section titles M Delete any section Entitled Endorsements Such a section may not be included in the Modified Version N Do not retitle any existing section to be Entitled Endorsements or to conflict in title with any Invariant Section O Preserve any Warranty Disclaimers If the Modified Version includes new front matter sections or appendices that qualify as Secondary Sections and contain no material copied from the Document you may at your option designate some or all of these sections as invariant To do this add their titles to the list of Invariant Sections in the Modified Version s license notice These titles must be distinct from any other section titles You may add a section Entitled Endorsements provided it contains nothing but endorsements of your Modified Version by various parties for example statements of peer review or that the text has been approved by an organization as the authoritative definition of a standard You may add a passage of up to five words as a Front Cover Text and a passage of up to 25 words as a Back Cover Text to the end of the list of Cover Texts in the Modified Version Only one passage of Front C
215. rez 153 204 Universidad Complutense de Madrid Datalog Educational System tapi test_tapi Answer Regular Remarks This command is used to test the current connection Example Input tapi test_tapi Output Ssuccess e Command tapi open_db db Arguments db Database connection name Not delimited Answer Regular Remarks This command is used to open an ODBC connection cf Section 5 13 2 Example Input tapi open_db test Output Ssuccess e Command tapi close_db Answer Regular Remarks This command is used to close the current ODBC connection cf Section 5 13 2 Example Input tapi close_db Output S success e Command tapi current_db Answer Fernando Sdenz P rez 154 204 Universidad Complutense de Madrid Datalog Educational System Two lines the first one containing the current ODBC connection name and the second one the external DBMS cf Section 5 13 2 Remarks This command is used to get the current ODBC connection name cf Section 5 13 2 Example Input assuming that the ODBC connection test is already opened tapi current_db Output test access Command tapi relation_exists relation_name Arguments relation_name Relation table view or predicate name which must be enclosed between delimiters if needed Answer Boolean Remarks This command returns true if the given relation exists and false otherwise
216. ring of up to n characters VARCHAR2 n Oracle s variable length string of up ton characters VARCHAR variable length string of up to the maximum length of the underlying Prolog atom STRING As VARCHAR CHARACTER VARYING n equivalent to the former INT INTEGER equivalent to the former SMALLINT NUMERIC p d a total of p digits where d of those are in the decimal place REAL Fernando Sdenz P rez 76 204 Universidad Complutense de Madrid Datalog Educational System DOUBLE PRECISION equivalent to the former S err with precision of at least n digits ee four digit year month and day vie hours minutes and seconds i rene combination of date and time ColumnConstraint NOT NULL PRIMARY KEY l UNIQUE CANDIDATE KEY REFERENCES TableName Att TableConstraints TableConstraint TableConstraint TableConstraint UNIQUE Att Att CANDIDATE KEY Att Att PRIMARY KEY Att Att FOREIGN KEY Att Att REFERENCES TableName Att Att RelationName is a user identifier for naming tables views and aliases TableName is a user identifier for naming tables ViewName is a user identifier for naming views Att is a user identifier for naming relation attributes LEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEES DML Data Manipulation Language statements LEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEES DMLstmt
217. rmal behaviour of current relational database systems e In contrast to Datalog built in identifiers are not case sensitive This conforms to the normal behaviour of current relational database systems 4 2 4 Data Definition Language This part of the language deals with creating or replacing and dropping tables and views There is no provision for updating the schema which can be consulted with the command dbschema 4 2 4 1 Creating Tables The first form of this statement is as follows CREATE OR REPLACE TABLE TableName Columnl1 Typel ColumnConstrainti1 ColumnN TypeN ColumnConstraintN TableConstraints This statement defines the table schema with name TableName and column names Column1 ColumnN with types Typel TypeN respectively If the optional clause OR REPLACE is used the table is dropped if existed already deleting all of its tuples A second form of this statement allows to create a table with the same schema of an existing table following SQL standard optional feature T171 CREATE TABLE TableName LIKE ExistingTableName Parentheses are not mandatory though This version copies the complete schema including all integrity constraints both predefined and user defined There is provision for several column constraints e NOT NULL Existency constraint forbiding null values PRIMARY KEY Primary key constraint for only one column e UNIQUE Uniqueness constraint for only one column
218. rom Datalog as well please remember to enable duplicates to get the expected result DES SQL gt datalog DES gt duplicates on Fernando Sdenz P rez 96 204 Universidad Complutense de Madrid Datalog Educational System Info Duplicates are on DES gt s X t X Info Processing answer X s X t X answer 1 answer 1 Info 2 tuples computed This is equivalent to the following SQL statement DES gt select s a from s t where s a t a answer a varchar gt answer 1 answer 1 Info 2 tuples computed However whilst the former has been processed by the Datalog engine the latter has been processed by the external RDBMS So some complex SQL statements might be more efficiently processed by the external RDBMS Duplicates are relevant in a number of situations For instance consider the following where duplicates are initially disabled DES gt group_by v X Y X Y C count Info Processing answer X Y C group_by v X Y X Y C count answer 1 1 1 answer 1 2 1 Info 2 tuples computed Although there are a couple of tuples for each group see the table contents above only one is returned in the count because they are indistinguishable in a set Now if duplicates are allowed we get the expected result DES gt duplicates on Info Duplicates are on DES gt group_by v X Y X Y C count Info Processing answer
219. rom origin to destination Figure 1 Paths in a Graph The file paths dl contains the following Datalog code which can be consulted with c paths Paths in a Graph edge a b edge a c edge b a edge b d path X Y path X Z edge Z Y path X Y edge X Y The query path X Y yields the following answer path a a path a b path a c path a d path b a path b b path b c path b d Info 8 tuples computed The file paths sql contains the SQL counterpart code which can be executed with process paths sql create table edge origin destination insert into edge values a b insert into edge values a c insert into edge values b a insert into edge values b d create view paths origin destination as with recursive path origin destination as 11 Adapted from TS86 Fernando Sdenz P rez 170 204 Universidad Complutense de Madrid Datalog Educational System select from edge union select path origin edge destination from path edge where path destination edge origin select from path So you can get the same answer as before with the SQL statement DES SQL gt select from paths answer paths origin paths destination gt answer a a answer a b answer a c answer a di answer b a answer b b answer b c answer b d Info 8 tuples computed Another shorter formulation is allowed in DES wi
220. ropping an inexistent table does not raise an error Example DROP TABLE t 4 2 44 Dropping Views DROP VIEW ViewName This statement drops the view with name ViewName deleting all of its tuples whether they were inserted with INSERT or with the command assert and rules which might have been added via assert Other tuples or rules asserted with the command assert are not deleted Example DROP VIEW v 4 2 4 5 Renaming Tables RENAME TABLE TableName TO NewTableName This non standard statement following IBM DB2 allows to change the name of table TableName to NewTableName Foreign keys referring to this table are modified accordingly Also views including referenes to this table are modified to refer to the new name 4 2 4 6 Renaming Views RENAME VIEW ViewName TO NewViewName This non standard statement following IBM DB2 allows to change the name of view ViewName to NewViewName Also views including references to this view are modified to refer to the new name 4 2 4 7 Dropping Databases DROP DATABASE This statement drops the current database dropping all tables views and rules this includes Datalog rules and constraints that may have been asserted or consulted It behaves exactly as the command abolish Example DROP DATABASE 4 2 5 Data Manipulation Language This part of the language deals with inserting and deleting tuples from tables There is no provision for updating tuples 4 2 5 1 In
221. rovided that SWI is the variable which holds the absolute filename of the SWI Prolog executable 2 2 3 Starting DES from a Prolog interpreter Besides the methods just described you can start DES from a Prolog interpreter disregarding the OS and platform first changing to the distribution directory and then submitting KE KE KE 3 des Or better if the system does support it ensure_loaded des If the system does not start by itself then type start Getting Started Whichever method you use to start DES a script batch file or shortcut as described in Section 2 2 you get the following Fernando Sdenz P rez 15 204 Universidad Complutense de Madrid Datalog Educational System kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk DES Datalog Educational System v 3 0 Type help for help about commands Fernando Saenz Perez c 2004 2012 GPD DISIA UCM Please send comments questions etc to fernan sip ucm es Web site http des sourceforge net FF FF FF FH HF HF HF FF FF FF HF HF HF HF HF This program comes with ABSOLUTELY NO WARRANTY is free software and you are welcome to redistribute it under certain conditions Type license for details KEKKKKKKKKKKKKKKKKKEKKKKKKEKKEKKEKKKKKKKKKKKKKKKKKKKKKKKKKKKK DES gt This last line DES gt is the DES system prompt which allows you to write Datalog SQL and Relational Algebra RA queries c
222. rt into t values 1 Not rejected as it is not monitored by DES Info 1 tuple inserted DES does monitor the following assertion as it is directed to the local database DES gt assert t 1 Error Primary key violation t a when trying to insert t l1 Error Asserting rules due to integrity constraint violation DES gt use_db Sdes When the current database is the local database des the external table t is not visible So the following fact is asserted in the local database DES gt insert into t values 1 Info 1 tuple inserted Any other attempt to assert the same fact t 1 is rejected DES gt assert t 1 Error Primary key violation t a when trying to insert t 1 Error Asserting rules due to integrity constraint violation The following would also go to the local database DES gt insert into t values 1 Error Primary key violation t a when trying to insert CU Error Asserting rules due to integrity constraint violation Info 0 tuples inserted Fernando Sdenz P rez 101 204 Universidad Complutense de Madrid Datalog Educational System Finally any persisted predicate see Section 5 2 which has attached constraints is checked for its consistency irrespective of the external database it is stored Also any of the supported constraints can be attached to persistent predicates therefore providing a high expressivity and declarative consistency level 5 1 9 Caveats and Limitati
223. rving this relation i e state n n n n Farmer takes Goat to south shore gt state s n s n Farmer returns to north shore gt state n n s n Farmer takes Wolf to south shore gt Fernando Sdenz P rez 178 204 Universidad Complutense de Madrid Datalog Educational System state s s s n Farmer takes Goat to north shore 3 state n s n n Farmer takes Cabbage to south shore gt state s s n s Farmer returns to north shore 23 state n s n s Farmer takes Goat to south shore gt state s s s s Final safe state Observe that there is two states in the relation state 4 which do not form part of the previous path state s n s s state n n n s These states come from another possible path 14 state n n n n Farmer takes Goat to south shore gt state s n s n Farmer returns to north shore gt state n n s n Farmer takes Cabbage to south shore gt state s n s s Farmer takes Goat to north shore 3 state n n n s Farmer takes Wolf to south shore 3 state s s s n Farmer takes Goat to north shore gt state s s n s Farmer returns to north shore 23 state n s n s Farmer takes Goat to south shore gt state s s s s Final safe state 6 9 Paradoxes files russell d1 sql rail When negation is used we can find paradoxes such as the Russell s paradox the barber in a town shaves every person who does not shave himself shown in the next example please note that this exa
224. s Info No tables Info No views Info No integrity constraints If you want to completely remove a predicate even its persistent representation you can use the command abolish as in DES gt abolish p DES gt dbschema Info Database Sdes Info No tables Info No views Info No integrity constraints DES gt listing p Info 0 rules listed DES gt use_db mysql DES gt dbschema mysql p Info Database mysql Error No table or view found with name p Also dropping the SQL view corresponding to a predicate removes persistency as in DES gt persistent t a int mysql DES gt dbschema Info Database Sdes Info No tables Info View s t a number integer Defining SQL statement CREATE VIEW t a AS SELECT ALL FROM t_des_table Info No integrity constraints DES gt drop view t DES gt dbschema Info Database Sdes Info No tables Info No views Info No integrity constraints 5 2 7 Schema and Data Visibility The default database DDB is called des and it contains metadata of each predicate for which either a type assertion or a SQL table creation statement has been Fernando S enz P rez 111 204 Universidad Complutense de Madrid Datalog Educational System issued If one makes a predicate persistent in an external database EDB its metadata as well as its data is visible both to DDB and EDB The following session illustrates this DES gt Zuse db
225. s hypothetical queries can be used For example and following the above system session DES gt assume select 3 1 in path a b select from path answer path a number integer path b number integer gt answer 1 1 answer 1 2 answer 1 3 answer 2 1 answer 2 2 answer 2 3 answer 3 1 answer 3 2 answer 3 3 Info 9 tuples computed This example also shows that DES is able to compute more queries than an RDBMS For instance neither MS SQL Server nor DB2 allow cycles in the above path definition This is not the most important limitation of recursion in current RDBMSs note that stratified recursion is not supported for more than one stratum This means that recursive SQL queries involving EXCEPT NOT IN aggregates are not allowed in current RDBMSs such as SQL Server and DB2 Another limitation is linear recursion the above rules cannot be expressed in a RDMBS s SQL as there are several recursive calls To name another UNION ALL is enforced in those SQLs so that just UNION is not allowed For instance the following query is rejected in any current commercial RDBMS but accepted by DES DES gt duplicates on DES gt multiline on DES gt CREATE TABLE edge a int b int DES gt INSERT INTO edge VALUES 1 2 Info 1 tuple inserted DES gt INSERT INTO edge VALUES 2 3 Info 1 tuple inserted DES gt INSERT INTO edge VALUES 1 3 Info 1 tuple inserted DES gt persistent edge a int
226. se de Madrid Datalog Educational System If a section in the Document is Entitled Acknowledgements Dedications or History the requirement section 4 to Preserve its Title section 1 will typically require changing the actual title 9 TERMINATION You may not copy modify sublicense or distribute the Document except as expressly provided under this License Any attempt otherwise to copy modify sublicense or distribute it is void and will automatically terminate your rights under this License However if you cease all violation of this License then your license from a particular copyright holder is reinstated a provisionally unless and until the copyright holder explicitly and finally terminates your license and b permanently if the copyright holder fails to notify you of the violation by some reasonable means prior to 60 days after the cessation Moreover your license from a particular copyright holder is reinstated permanently if the copyright holder notifies you of the violation by some reasonable means this is the first time you have received notice of violation of this License for any work from that copyright holder and you cure the violation prior to 30 days after your receipt of the notice Termination of your rights under this section does not terminate the licenses of parties who have received copies or rights from you under this License If your rights have been terminated and not permanently reinstat
227. serting Tuples Fernando Sdenz P rez 65 204 Universidad Complutense de Madrid Datalog Educational System INSERT INTO TableName VALUES Ctel CteN This statement inserts into the table TableName a tuple built with the values Ctel CteN A value for each column in the table has to be provided here N is the number of columns of TableName Example INSERT INTO t VALUES 1 1 Another form of the INSERT statement allows to inserting tuples which are the result set from a SELECT statement INSERT INTO TableName SQLStatement This statement inserts into the table TableName as many tuples as returned by the SQL statement SQLStatement This statement has to return as many columns as the columns of TableName Examples INSERT INTO t SELECT FROM s You can also insert tuples coming directly or indirectly from a table as in INSERT INTO t SELECT FROM t For testing the new duplicated contents of t you have to use listing t instead of a SELECT since this statement always returns a set no duplicates when duplicates are disabled cf Section 4 1 9 4 2 5 2 Deleting Tuples DELETE FROM TableName This statement deletes all the tuples of the table TableName It does not delete production rules asserted via assert Example DELETE FROM t Another form of the DELETE statement allows to deleting tuples which fulfil a given condition DELETE FROM TableName WHERE Condition This statement deletes fro
228. ses and an extended SQL debugger following CGS12a The first one motivates raising the major version as it opens a brand new scenario for several reasons First predicates are no longer limited by available memory instead persisted predicates are using as much secondary storage as needed and provided by the underlying external database Second processing is directed to the external database for those rules that can be projected and to the deductive engine for rules that can not This way one can take advantage of the external database performance and scalability Third queries which are not possible in an external database can be solved by the deductive engine So one can extend external database expressiveness with the added features in DES Finally as several ODBC connections are allowed at a time different predicates can be made persistent in different DMBSs which allows for interoperability among external relational engines and the local deductive engine therefore enabling business intelligence applications See Section 5 2 for details With respect to the new SQL Debugger version which is a new implementation it also now accept user information about wrong and missing tuples allowing to focus the questions directed to the user on more precise error sources therefore preventing many questions involving large sets of data see Section 5 9 for details New ports to SICStus Prolog 4 2 1 and SWI Prolog 6 0 2 have been provided License
229. sider the following session where it can be seen that the system monitors type constraints in both Datalog and SQL queries DES gt type p int string DES gt assert p a b Error Type mismatch p 1 number integer vs string char _6372 p 1 number integer 2 string varchar DES gt assert p 1 a DES gt p X Y Fernando Sdenz P rez 49 204 Universidad Complutense de Madrid Datalog Educational System p 1 a Info 1 tuple computed DES gt select from p answer p 1 p 2 gt answer 1 a Info 1 tuple computed DES gt insert into p values a b Error Type mismatch p 1 number integer vs string char _6937 p 1 number integer 2 string varchar Info 0 tuples inserted Note that columns with automatically given names can be accessed from a SQL statement but enclosed as special user identifiers ISO delimiters double quotes supported by Oracle and SQL Server are supported as well as other vendor specific delimiters MS Access square brackets and MySQL back quotes Otherwise an error is raised DES gt select 1 from p Error Input processing error DES gt select 1 from p answer p 1 gt answer 1 Info 1 tuple computed A relation already defined is checked for consistency when trying to assert a new type constraint DES gt assert t 1 DES gt assert t a DES gt type t int Error No type tupl
230. sitive goal The command list_et shows the current state of the extension table both for answers and calls already obtained by solving one or more queries incidentally recall that you can focus on the contents of the extension table for a given predicate cf Section 5 13 4 This command is useful for the user when asking for the meaning of relations and for the developer for examining the last calls being performed Before executing any query the extension table is empty after executing a query at least the call is not empty Also the extension table is empty after the execution of a temporary view II The extension table contains the calls made during the last fixpoint iteration see next section for details the calls are cleared before each iteration whereas the answers are kept The command clear_et clears the extension table contents both for calls and answers 8 For a complementary understanding of this section the reader is advised to read Diet87 9 A term T1 subsumes a term T2 if T1 is more general than T2 and both terms are unifiable Eg p X subsumes p a Z p X Y subsumes p U V p X Y subsumes p U U but p U U neither subsumes p a b nor p X Y 10 The contents of the extension table in this case should be restored instead of being cleared left for further improvements Fernando S enz P rez 164 204 Universidad Complutense de Madrid Datalog Educational System 5 16 2 Fixpoint Computa
231. statement i e tuples intensionally assumed As an example let s suppose that the relation flight is as previously defined and a view connect that displays locations connected by direct flights DES gt CREATE VIEW connect origin destination AS SELECT origin destination FROM flight DES gt SELECT FROM connect answer connect origin string varchar connect destination string varchar gt answer Lon pi answer mad par answer par ny Info 3 tuples computed Fernando S enz P rez 73 204 Universidad Complutense de Madrid Datalog Educational System Then if we assume that connections are allowed with transits we can submit the following hypothetical query note that the assumed SQL statement is recursive DES gt ASSUME SELECT flight origin connect destination FROM flight connect WHERE flight destination connect origin IN connect origin destination SELECT FROM connect answer connect origin string varchar connect destination string varchar gt answer lon ny answer mad ny answer mad par answer par ny Info 4 tuples computed In addition to this one can use a WITH statement instead of an ASSUME statement by simply stating an existing relation in the definition of the local view For instance for the last example we can write DES gt WITH connect origin destination AS SELECT flight origin connect destination FROM flight connect WHERE f
232. stem session shows an example DES gt type p a int b string DES gt nn p a The list of column names specifies the columns for which null values are not allowed Thus trying to assert a tuple such as the following will raise an error DES gt assert p null Error Not null violation p a Subsequent existency constraints are allowed for the same predicate and arity the last declaration is the one to persist overriding previous declarations for such predicate 4 1 14 3 Primary Key A primary key constraint specifies that no two tuples have the same values for a given set of columns Next a system session illustrates the use of a primary key assertion DES gt type p a int b string Fernando Sdenz P rez 51 204 Universidad Complutense de Madrid Datalog Educational System DES gt pk p a Primary key constraints are trivially satisfied when duplicates are disabled as relations are considered as sets irrespective of the current database instance that may contain duplicates for the arguments in the primary key Several primary key declarations are allowed for the same predicate and arity the last declaration is the one to persist overriding previous type declarations for such predicate DES gt pk p a DES gt pk p c Error Unknown column c DES pk p a a A relation already defined with facts or rules is checked for consistency when trying to assert a
233. such names between separators For instance file or directory names can contain blanks for Windows users and you neither need to use double quotes nor are allowed to use them Since commands are submitted with a preceding slash they are only recognized as commands in this way Therefore you can use command names for your relation names without name clashes When consulting Datalog files filename resolution works as follows e If the given filename ends with d1 DES tries to load the file with this absolute or relative filename e If the given filename does not end with d1 DES firstly tries to load a file with d appended to the end of the filename If such a file is not found it tries to load the file with the given filename In command arguments when applicable you can use relative or absolute pathnames In general you can use a slash as a directory delimiter but depending on the platform you can also use the backslash Also it might be needed to enclose pathnames between single quotes See Section 4 1 2 for information about DES queries Some commands are labelled with TAPI enabled which means that they can be submitted to the textual application programming interface TAPI There is additional information for such commands in Section 5 14 2 Next commands are described where italics indicate a parameter which must be supplied by the user Square brackets indicate an optional keyword or parameter except
234. swer a2 1 Info 2 tuples computed DES RA gt Renaming DES RA gt select al a lt a2 a rename al a ail product rename a2 a a answer al a string varchar a2 a string varchar gt answer al a2 answer al a3 answer a2 a3 Info 3 tuples computed DES RA gt Duplicate elimination DES RA gt duplicates off Info Duplicates are already disabled DES RA gt project a c answer c a string varchar gt Fernando Sdenz P rez 25 204 Universidad Complutense de Madrid Datalog Educational System answer al answer a2 Info 2 tuples computed DES RA gt duplicates on DES RA gt project a c answer c a string varchar gt answer al answer al answer a2 Info 3 tuples computed DES RA gt distinct project a c answer c a string varchar gt answer al answer a2 Info 2 tuples computed 3 4 Prolog Mode This mode is enabled via the command prolog and goals are sent to the Prolog processor Assuming that the file relop dl has been already consulted let s consider the following example DES Prolog gt projection X projection al type for more solutions lt Intro gt to continue projection al type for more solutions lt Intro gt to continue projection a2 type for more solutions lt Intro gt to continue no DES Prolog gt datalog projection X projection al projection a2 I
235. t a query with two arguments as follows X is Expression where X is a variable or anumber and Expression is an arithmetic expression built from numbers variables built in arithmetic operators constants and functions mainly following ISO for Prolog they are labelled if so in the listings below Availability of arithmetic built ins mainly depend on the underlying Prolog system binary distributions cope with all the listed built ins At evaluation time the expression must be ground i e its variables must be bound to numbers or constants otherwise problems as stated in the previous section may arise Evaluating the above query amounts to evaluate the arithmetic expression according to the usual arithmetic rules which yields a number integer or float and X is bound to this number if it is a variable or tested its equivalence if it is a number Precision depends on the underlying Prolog system Arithmetic built ins have meaning only in the second argument of is they cannot be used elsewhere For example DES gt X is sqrt 2 1 4142135623730951 is sqrt 2 Info 1 tuple computed Here sqrt 2 is an arithmetic expression that uses the built in function sqrt square root But DES gt sqrt 2 is sqrt 2 Fernando Sdenz P rez 88 204 Universidad Complutense de Madrid Datalog Educational System raises an input error because an arithmetic expression can only occur as the right argument of is A
236. t also for Theta join operator outer join operators duplicate elimination distinct operator grouping group_by operator recursive queries and renaming operator this avoids to resort to building new relations with the assignment operator although it is supported too RAstmt SELECT WhereCondition B rel Selection sigma PROJECT SelectExpressionList RArel Projection pi RENAME Schema RAre1 Renaming rho EE B rel Duplicate elimination or PRODUCT RArel Cartesian Product E UNION RArel Set union EE DIFFERENCE B rel Set difference en INTERSECT RArel Set intersection Ge NJOIN RArel Natural join AT ZJOIN WhereCondition RArel Zeta join EE LJOIN WhereCondition B rel Left outer join hed RJOIN WhereCondition RArel Right outer join ee FJOIN WhereCondition RArel Full outer join GROUP_BY Atts SelectExpressionList HavingCondition RArel Grouping RArel RAstmt Relation View definition assignment statement RAview Schema RAstmt Relation Schema ViewName ViewName ColName ColName Fernando Sdenz P rez 86 204 Universidad Complutense de Madrid Datalog Educational System WhereCondition SelectExpressionList and HavingCondition are asin SQL grammar 4 4 Prolog Syntax of Prolog programs and goals is the same as for Datalog including all built in operators cf next Section but aggregates Notice that negation is written
237. t least one tuple in the result set of the view not verifying the where condition In queries containing aggregate functions this tuple either does not satisfy either the where condition or the having condition Set operations are also allowed in both PTC and NTC generation It is possible to obtain a test case which is both positive and negative at the same time thus achieving predicate coverage with respect to the where and having clauses in the sense of AO08 We will call these tests PNTCs For instance consider the following system session DES SQL gt create table t a int primary key DES SQL gt create view v a as select a from t where a 5 DES SQL gt test_case v Info Test case over integers t 5 t 5 The test case t 5 t 4 is a PNTC However a PNTC is not always possible to be generated For instance it is possible for the following view to generate both PTCs and NTCs but no PNTC create view v a as select a from t where a 1 and not exists select a from t where a lt gt 1 The only one PTC for this view is t 1 modulo duplicates There are many NTCs as e g t 2 and t 1 t 2 The command test_case View Options allows two kind of options first to specify which class of test case is to be generated all PNTC the default option positive PTC or negative NTC The second option specifies an action the results are to be displayed via the option display default option add
238. t value of X and Y e max X Y Maximum Greatest value of X and Y e truncate X Truncate ISO Closest integer between X and 0 e float_integer_part X Integer part as a float ISO The same as float integer X e float_fractional_part X Fractional part as a float ISO Fractional part of X i e X float_integer_part X e round X Closest integer ISO Closest integer to X X has to be a float If X is exactly half way between two integers it is rounded up i e the value is the least integer greater than X e floor X Floor ISO Greatest integer less or equal to X X has to be a float e ceiling X Ceiling ISO Least integer greater or equal to X X has to be a float Fernando Sdenz P rez 91 204 Universidad Complutense de Madrid Datalog Educational System 4 5 5 Negation e not Query Stratified negation It stands for the complement of the relation Query w r t the meaning of the program i e closed world assumption See Sections 4 1 8 and 5 16 3 If Query is not an atom a new predicate defined by a head Head with relevant variables in Query is built and defined by the single rule Head Query Then not Head replaces not Query 4 5 6 Datalog Outer Joins e 154 LeftRelation RightRelation JoinCondition Left join It stands for the left outer join of the relations LeftRelation and relations RightRelation under the condition JoinCondition expressed as literals cf Section 4
239. tables DES SQL gt insert into a values al Info 1 tuple inserted DES SQL gt insert into a values a2 Info 1 tuple inserted DES SQL gt insert into a values a3 Info 1 tuple inserted DES SQL gt insert into b values bl Info 1 tuple inserted DES SQL gt insert into b values b2 Info 1 tuple inserted DES SQL gt insert into b values al Info 1 tuple inserted DES SQL gt insert into c values al b2 Info 1 tuple inserted DES SQL gt insert into c values al al Info 1 tuple inserted DES SQL gt insert into c values a2 b2 Info 1 tuple inserted DES SQL gt Testing the just inserted values DES SQL gt select from a answer a a gt answer al answer a2 answer a3 Info 3 tuples computed DES SQL gt select from b answer b b gt answer al answer b1 answer b2 Info 3 tuples computed DES SQL gt select from c answer c a c b gt answer al ali answer al b2 answer a2 b i Info 3 tuples computed DES SQL gt Projection DES SQL gt select a from c answer c a gt answer al answer a2 Fernando S enz P rez 20 204 Info 2 tuples computed DES SQL gt Selection DES SQL gt select a from a where a a2 answer a a gt answer a2 Info 1 tuple computed DES SQL gt Cartesian product DES SQL gt select from a b answer a a b b gt answer al ali answer al b1
240. tata E E E E 185 Fernando S enz P rez 6 204 Universidad Complutense de Madrid Datalog Educational System Ze COMELUD Dr CC o k ceeccesceeteicdeceesesscsevchtives oosveascovacusevsedudectecessvestudesonesiedsnseustede obsdecienchddacvencsees 820 Related WOK siceiccscccasssasivacsecccecssccseessstesescssvcdeshivadssevsecsevesebudsseevsedssdevsdessdevensedudstvectssuseces 8 1 D ductive Database Systems eerste 8 2 Technological E 9 Fut te Enhancements ensesinde overeen stire esrara a epee eases tra e ee 10 Caveats and Limitations isssscccsseasscstsoceseceosecseecisecsessdvecsvodssenssdesso cussed seoveseseecevdsavedasens TT Release Notes ieceiscssecssteccasetuctsisesssoecevescouetocese en cobededeueveds seos ecesiuededees Wtevsetdecvedsesbaceces 11 1 Version 3 0 of DES released on May 9th 2012 cesicusiescacsscrecerecticaareiiemetionnes TA Acknowledgements siinsesse naisara saiia Esie inss Appendix As E BAD 1 Eat dee eege eege Fernando Sdenz P rez 7 204 8 Universidad Complutense de Madrid Datalog Educational System 1 Introduction The Datalog Educational System DES is a free open source multiplatform portable Prolog based implementation of a deductive database system DES 3 0 is the current implementation which enjoys Datalog Relational Algebra and SQL query languages full recursive evaluation with memoization techniques full fledged arithmetic stratified negation duplicates an
241. teFilename Synonym type Filename cd Path Set the current directory to Path TAPI enabled cd Set the current directory to the directory where DES was started from TAPI enabled Fernando Sdenz P rez 142 204 Universidad Complutense de Madrid Datalog Educational System 5 13 6 5 13 7 pwd Display the absolute filename for the current directory TAPI enabled 1s Display the contents of the current directory in alphabetical order First files are displayed then directories Synonym dir 1s Path Display the contents of the given directory in alphabetical order It behaves as 1s Synonym dir Path shell Command Submit Command to the operating system shell Notes for platform specific issues o Windows users command exe is the shell for Windows 98 whereas cmd exe is the one for Windows NT 2000 2003 XP Vista 7 o Ciao users The environment variable SHELL must be set to the required shell o SICStus users Under Windows if the environment variable SHELL is defined it is expected to name a Unix like shell which will be invoked with the option c Command If SHELL is not defined the shell named by COMSPEC will be invoked with the option C Command o Windows and Linux Unix executable users The same note for SICStus is applied Synonyms s rm FileName Delete FileName from the file system Synonyms del Log log Display the current log file if any log Filen
242. th the following view definition create view path origin destination as select from select from edge union select path origin edge destination from path edge where path destination edge origin You can finally compare this with the RA formulation paths origin destination select true edge union project paths origin edge destination edge zjoin paths destination edge origin paths 6 3 Shortest Paths file spaths d1 sql ra Thanks to aggregate predicates one can code the following version of the shortest paths problem file spaths d1 which uses the same definition of edge as the previous example path X Y 1 edge X Y path X Y L path X Z L0 edge Z Y count edge A B Max Fernando S enz P rez 171 204 Universidad Complutense de Madrid Datalog Educational System LO lt Max L is LO 1 sp X Y L min path X Y Z 2Z L Note that the infinite computation that may raise from using the builtin is 2 is avoided by limiting the total length of a path to the number of edges in the graph The following query returns all the possible paths and their corresponding minimal distances DES gt sp X Y L sp a a 2 sp a b 1 sp a c 1 sp a d 2 sp b a 1 sp b b 2 sp b c 2 sp b d 1 Info 8 tuples computed Below is the SQL formulation for the same problem file spaths sq1 DES SQL gt create or replace view spaths origin destination
243. the user should the file exists already Constraints type nullability primary key candidate key functional dependency foreign key and user defined are also saved ODBC Database open_db Name Options Open and set the current ODBC connection to Name where Options user Username password Password This connection must be already defined at the OS layer TAPI enabled close_db Close the current ODBC connection TAPI enabled close_db Name Close the given ODBC connection TAPI enabled current_db Display the current ODBC connection name and DSN provider TAPI enabled show_dbs Display the open database connections TAPI enabled use_db Name Make Name the current ODBC connection TAPI enabled Debugging and Test Case Generation debug_datalog Goal Level Start the debugger for the basic goal Goal at predicate or clause levels which is indicated with the options p and c for Level respectively Default is p debug_sql View Options Debug a SQL view where Options trust_tables yes no trust_file FileName Defaults are trust tables and no trust file It might be needed to enclose FileName between single quotes trace_datalog Goal Order Trace a Datalog goal in the given order postorder or the default preorder Fernando Sdenz P rez 141 204 Universidad Complutense de Madrid Datalog Educational System 5 13 4 5 13 5 trace_sql View Order Trace a SQL view in the give
244. this the expected answer for view AnimalOwner y n m mT w wN a h yl y Info Buggy relation found CatsAndDogsOwner In this example tables have been trusted but it is also possible to ask the user for the validity of the involved tables in the debugging process via the command debug_sql Guest trust_tables no In this example session validity of table Owner would be asked to the user 5 9 1 Trusted Specifications In SQL the following scenario is very usual A set of correct views is updated to improve its efficiency The new set of views includes both new views and improved versions of some old views keeping their names and intended answers Sometimes the new usually more involved system no longer produces the expected results We allow to use the first reliable version which we call a trusted specification during the subsequent debugging session For instance let s consider that the user has corrected the former example which is now working properly Now suppose that in order to improve readability the set of views is changed by removing AnimalOwner adding instead a new view CatOrDogOwner and modifying LessThan6 and CatsAndDogsOwner which now make use of CatOrDogOwner Fernando Sdenz P rez 130 204 Universidad Complutense de Madrid Datalog Educational System Next the modified and new views Guest and NoCommonName remain the same this new version is located in file examples SQLDebugger pets2 sql are
245. till consistent w r t to the constraint However trying to add the second rule for path makes it infeasible so that it is rejected Now only 5 rules have been asserted If the file was not included the third fact for edge then it would be accepted as a valid tree Again trying to insert such a tuple after such a program is consulted raises an error DES gt assert edge d a Info Checking user defined integrity constraint over database path X X Info Computing predicate dependency graph Info Computing strata Error Integrity constraint violation ic X path X X Fernando S enz P rez 57 204 Universidad Complutense de Madrid Datalog Educational System Offending values in database ic a ic b ic d Observe that since the path relation is now complete all the nodes in the cycle are displayed a b and c The considered constraint is not yet enough to ensure a directed tree defined by edge facts Two conditions remain First a given node cannot have more than one incoming edge and second a tree must be a connected graph If the first condition is imposed it suffices for the second to check that the number of nodes is the number of edges plus 1 So DES gt assert node N edge N A edge A N Info Computing predicate dependency graph Info Computing strata Info Rule asserted DES gt count edge A B Es count node N Ns D is Ns Es D 1 Info Parsing query Inf
246. tion The tabling mechanism is insufficient in itself for computing all of the possible answers to a query The rationale behind this comes from the fact that the computed information is not complete when solving a given goal because it can use incomplete information from the goals in its defining rules these goals can be mutually recursive Therefore we have to ensure that we produce all the possible information by finding a fixpoint of the memo function First the call table is emptied in order to allow the system to try to obtain new answers for a given call preserving the previous computed answers Then the memo function is applied possibly providing new answers If the answer table remains the same as before after this last memo function application we are done Otherwise the memo function is reapplied as many times as needed until we find a stable answer table with no changes in the answer table The answer table contains the stable model of the query plus perhaps other stable models for the relations used in the computation of the given query The fixpoint is found in finite time because the memo function is monotonic in the sense that we only add new entries each time it is called while keeping the old ones Repeatedly applying the memo function to the answer table delivers a finite answer table since the number of new facts that can be derived from a Datalog program is finite recall that there are no compound terms such as s z
247. tion body Solving recursion in RA has been proposed as the application of a fixpoint operator to an RA expression see for instance Agra88 HA92 DES compiles RA expressions to Datalog programs and uses the fixpoint based deductive engine to solve them As an example of recursion in RA let s consider the following classic program for finding paths in a graph create table edge origin string destination string paths origin destination select true edge union project paths origin edge destination select paths destination edge origin edge product paths select true paths 4 3 3 RA Grammar Here terminal symbols are Parentheses commas semicolons single dots asterisks and apostrophes Other terminal symbols are completely written in capitals as SELECT However they are recognized by the parser in any letter case Percentage symbols start comments User identifiers must start with a letter and consist of letters and numbers otherwise a user identifier can be enclosed between quotation Fernando Sdenz P rez 85 204 Universidad Complutense de Madrid Datalog Educational System marks both square brackets and double quotes are supported and contain any characters Next RAstmt stands for a valid RA statement This grammar is built following Diet01 so that RA files read in WinRDBI a tool described in that book are also read in DES DES grammar extends WinRDBI grammar in providing suppor
248. tml Contact markus triska gmx at Installation Copy des e1 in the contributors web page to your home directory and add to your emacs load des adapt the following path as necessary setq des prolog file des systems swi des pl1 add to list auto mode alist d1 des mode Restart Emacs open a d1 file to load it into a DES process this currently only works with SWI Prolog If the region is active F1 consults the text in the region You can then interact with DES as on a terminal 8 Related Work There has been a high amount of work around deductive databases RU95 its interest delivered many workshops and conferences for this subject which dealt to several systems However to the best of our knowledge there is no a friendly system oriented to introducing deductive databases with several query languages to students Nevertheless on the one hand we can comment some representative deductive database systems On the other hand also some technological transfers to face real world problems Fernando Sdenz P rez 186 204 Universidad Complutense de Madrid Datalog Educational System 8 1 Deductive Database Systems 4QL MS11 is a recent development of a rule based database query language with negation allowed in bodies and heads of rules which is founded on a four valued semantics with truth values true false inconsistent and unknown It provides means for a uniform treatment of Open
249. topwatch Ss display_ stopwatch Display stopwatch list_persisted Display the persisted predicates TAPI enabled show_dbs Display the open database connections TAPI enabled Ss show_sql Display whether SQL statements which are sent to an external database are to be displayed show_sql Switch Enable or disable display of SQL statements which are sent to an external database on or off resp Fernando Sdenz P rez 190 204 Universidad Complutense de Madrid Datalog Educational System Zuse db Name Make Name the current ODBC connection TAPI enabled dbschema Connection Name Display the database schema for the given view or table name in the given connection license Display GPL and LGPL licenses If not found please visit http www gnu org licenses o New assertions persistent PredSpec Connection Make a predicate to persist on an external RDBMS via an ODBC connection PredSpec can be either the pattern PredName Arity or PredName Schema where Schema can be either ArgNamel1 ArgNameN or ArgNamel Typel ArgNameN TypeN If a connection name is not provided the current open database is used o Binary flags in commands are no longer case sensitive o New port to SICStus Prolog 4 2 1 This release fixes in particular some issues with ODBC connections exceptions about misencoded string in non ASCII ODBC messages and incorrect handling of SQL_BIGINT and related types o New port to SWI
250. tring varchar gt answer al answer a2 answer a3 Info 3 tuples computed DES RA gt select true b answer b b string varchar gt answer al answer b1 answer b2 Info 3 tuples computed DES RA gt select true c answer c a string varchar c b string varchar gt answer al al answer al b2 answer a2 b2 Info 3 tuples computed DES RA gt Projection DES RA gt project a c answer c a string varchar gt answer al answer a2 Info 2 tuples computed DES RA gt Selection DES RA gt select a a2 a answer a a string varchar gt answer a2 Info 1 tuple computed DES RA gt Cartesian product DES RA gt a product b answer a a string varchar b b string varchar gt answer al al answer al bl answer al b2 answer a2 al Fernando S enz P rez 23 204 Universidad Complutense de Madrid Datalog Educational System answer ai DI answer a2 b2 answer ai ali answer ai DI answer ai b i Info 9 tuples computed DES RA gt Theta Join DES RA gt select a a b b a product b answer a a string varchar b b string varchar gt answer al al Info 1 tuple computed DES RA gt a zjoin a a b b b answer a a string varchar b b string varchar gt answer al al Info 1 tuple computed DES RA gt Natural Inner Join DES RA gt a njoin cC answer a a string varchar c b
251. ts possible direct flights between locations and travel represents possible connections by using one or more direct flights Both include flight time By querying the relation travel we get Fernando Sdenz P rez 72 204 Universidad Complutense de Madrid Datalog Educational System DES gt select from travel answer travel origin string varchar travel destination string v archar travel time number float gt answer lon ny 9 0 answer mad ny 11 5 answer mad par 1 5 answer par ny 10 0 Info 4 tuples computed Now if we assume there is a tuple flight mad lon 2 0 we can query the database with this assumption with the following query with multi line input enabled DES gt ASSUME SELECT mad lon 2 0 IN flight origin destination time SELECT FROM travel answer travel origin string varchar travel destination string v archar travel time number float gt answer lon ny 9 0 answer mad lon 2 0 answer mad ny 11 0 answer mad ny 11 5 answer mad par 1 5 answer par ny 10 0 Info 6 tuples computed Note that the SELECT statement following the keyword ASSUME simply stands for the construction of a single tuple for table flight such statement can be otherwise stated as SELECT mad lon 2 0 FROM dual where dual is the built in table described in Section 4 2 6 1 2 In addition not only tuples can be extensionally assumed but any SQL DQL
252. tt RelationName Att RelationName ArithmeticExpression DOLstmt RenamedExpression UnrenamedExpression AS Identifier ArithmeticExpression Op1 ArithmeticExpression ArithmeticExpression Op2 ArithmeticExpression ArithmeticFunction ArithmeticExpression ArithmeticExpression Number Att RelationName AC ArithmeticConstant DQLstmt Op1 se EC Op2 J vem 1 1 JN I lt lt gt gt ArithmeticFunction sgqrt 1 1ln 1 log 1 log 2 sin 1 cos 1 tan 1 cot 1 asin 1 acos 1 atan 1l acot 1 abs 1 float 1 integer 1 sign 1 gced 2 min 2 max 2 truncate 1 float_integer_part 1 float_fractional_part 1 round 1 floor 1 ceiling 1 Aggregate Functions The argument may include a prefix distinct for all but min and max avg 1 count 1 count 0O max 1 min 1 sum 1 times 1 Fernando Sdenz P rez 79 204 7 Universidad Complutense de Madrid Datalog Educational System ArithmeticConstant pi e Rels Rel Rel Rel UnrenamedRel RenamedRel UnrenamedRel TableName ViewName DQLstmt JoinRel RenamedRel rs UnrenamedRel AS Identifier JoinRel Rel NATURAL JoinOp Rel JoinCondition JoinOp INNER JOIN LEFT OUTER JOIN RIGHT OUTER JOIN FULL OUTER JOIN JoinCondition ON WhereCondition USING Atts WhereCo
253. turn the maximum value for Variable ignoring nulls 4 5 7 2 Group_by Predicate e group_by Query Variables GroupConditions Solve GroupConditions in the context of Query building groups w r t the possible values the variables in the list Variables This list is specified as a Prolog Fernando Sdenz P rez 92 204 Universidad Complutense de Madrid Datalog Educational System list i e a sequence of comma separated values enclosed between brackets If this list is empty there is only one group the answer set for Query The goal GroupConditions may contain expressions including aggregate functions 4 5 7 3 Aggregate Predicates e count Query Variable Result Count in Result the number of tuples in the result set for the query Query so that Variable is a variable of Query an attribute of the result relation set and this attribute is not null It returns 0 if no tuples are found in the result set e count Query Result Count in Result the total number of tuples in the result set for the query Query disregarding whether they contain nulls or not It returns 0 if no tuples are found in the result set e sum Query Variable Result Sum in Result the numbers in the result set for the query Query and the attribute Variable which should occur in Query Nulls are simply ignored e times Query Variable Result Compute in Result the product of all the numbers in the result set for the query Query and the attribute Variable wh
254. tware Testing Cambridge University Press 2008 F Arni K Ong S Tsur H Wang and C Zaniolo The deductive database system LDL TPLP 3 1 61 94 2003 F Bueno D Cabeza M Carro M Hermenegildo P Lopez Garcia and G Puebla The Ciao Prolog system Reference manual School of Computer Science Technical University of Madrid UPM 1997 http www clip dia fi upm es M Becker C Fournet and A Gordon Design and Semantics of a Decentralized Authorization Language In CSF 07 Proceedings of the 20th IEEE Computer Security Foundations Symposium pages 3 15 Washington DC USA 2007 IEEE Computer Society M L Barja N W Paton A Fernandes M H Williams A Dinn An Effective Deductive Object Oriented Database Through Language Integration In Proc of the 20 VLDB Conference 1994 Caballero R A declarative debugger of incorrect answers for constraint functional logic programs in WCFLP 05 Proceedings of the 2005 ACM SIGPLAN workshop on Curry and functional logic programming 2005 pp 8 13 A Cali G Gottlob and T Lukasiewicz Datalog a unified approach to ontologies and integrity constraints In ICDT 09 Proceedings of the 12th International Conference on Database Theory pages 14 30 New York NY USA 2009 ACM R Caballero Y Garcia Ruiz and F Sdenz P rez Towards a Set Oriented Calculus for Logic Programming Programaci n y Lenguajes P Lucio y F Orejas
255. uding the ending dot to continue Upon exceptions the extension table is cleared and stratification is recomputed Note that the latter computation may take a long time if there are multiple tables and views typically in opened ODBC connections for DBMS s as Oracle and SQL Server 5 13 Commands The input at the prompt Oe commands or queries must be written in a line i e without carriage returns although it can be broken by the DES console due to space limitations and can end with an optional dot Commands are issued by preceding the command with a slash at the DES system prompt Command arguments are not a comma separated list enclosed between brackets as usual but they simply occur separated by at least one blank This enables short typing Command names and binary flags on off switches are not case sensitive Fernando Sdenz P rez 137 204 Universidad Complutense de Madrid Datalog Educational System Ending dots are considered as part of the argument wherever they are expected For instance ed behaves as ed this command changes the working directory to the parent directory In this last case the final dot is not considered as part of the argument The command 1s_ shows the contents of the working directory whereas 1s shows the contents of the parent directory which behaves as 1s BK Filenames and directories can be specified with relative or absolute names There is no need of enclosing
256. uest 3 Robin Scott 2 Guest 4 Tom Cohen Input Is this the expected answer for view Guest y n m mT w wN a h n n Info view NoCommonName is nonvalid w r t the trusted file Info view LessThan6 is valid w r t the trusted file Info view CatsAndDogsOwner is nonvalid w r t the trusted file Info Debugging view CatsOrDogsOwner 1 CatsOrDogsOwner 1 Kitty cat 2 CatsOrDogsOwner 1 Wilma dog Fernando Sdenz P rez 131 204 Universidad Complutense de Madrid Datalog Educational System CatsOrDogsOwner 2 Lucky dog CatsOrDogsOwner 2 Wilma cat CatsOrDogsOwner 3 Oreo cat CatsOrDogsOwner 3 Rocky dog CatsOrDogsOwner 4 Chelsea dog zl Oh LD B W l Input Is this the expected answer for view CatsOrDogsOwner y n m mT w wN a h yl Info Buggy view found CatsAndDogsOwner Here the debugger traverses the computation tree as before but the user is not asked for views in the set of trusted views and the erroneous view is caught with only one final check compared to the four checks that would be needed otherwise The debugger detects that the new version of CatsAndDogsOwner is erroneous 5 9 2 Missing and Wrong Tuples The debugger also allows the user to specify the error type indicating if there is either a missing answer a tuple was expected but it is not in the result or a wrong answer the result contains an une
257. ultiauthor Collaboration or MMC contained in the site means any set of copyrightable works thus published on the MMC site CC BY SA means the Creative Commons Attribution Share Alike 3 0 license published by Creative Commons Corporation a not for profit corporation with a principal place of business in San Francisco California as well as future copyleft versions of that license published by that same organization Fernando Sdenz P rez 199 204 Universidad Complutense de Madrid Datalog Educational System Incorporate means to publish or republish a Document in whole or in part as part of another Document An MMC is eligible for relicensing if it is licensed under this License and if all works that were first published under this License somewhere other than this MMC and subsequently incorporated in whole or in part into the MMC 1 had no cover texts or invariant sections and 2 were thus incorporated prior to November 1 2008 The operator of an MMC Site may republish an MMC contained in the site under CC BY SA on the same site at any time before August 1 2009 provided the MMC is eligible for relicensing ADDENDUM How to use this License for your documents To use this License in a document you have written include a copy of the License in the document and put the following copyright and license notices just after the title page Copyright C YEAR YOUR NAME Permission is granted to copy distribute a
258. use one can follow the execution of a goal via the SLD resolution tree and use the four port debugging approach Datalog stems from logic programming and Prolog in particular and it can be also understood as a subset of Prolog However its operational behaviour is quite different since the outcome of a query represents all the possible resolutions instead of a single one as in Prolog In addition tabling cf Section 5 4 and program transformations due to outer joins aggregates simplifications disjunctions make tracing cumbersome Similarly SQL represents a true declarative language which is even farthest from its computation procedure than Prolog Indeed the execution plan for a query include transformations considering data statistics to enhance performance These query plans are composed of primitive relational operations such as Cartesian product and specialized operations for which efficient algorithms have been developed containing in general references to index usage Therefore instead of following a more imperative approach to tracing here we focus on a naive declarative approach which only take into account the outcomes at some program points This way the user can inspect each point and decide whether its outcome is correct or not This approach will allow to examine the syntactical graph of a query which possibly depends on other views or predicates SQL or Datalog resp This graph may be cyclic when recursive
259. uthor wishes to thank the Clip group for providing their free Ciao system and in particular to F Bueno and J Correas for his help in porting DES to the Ciao system Also thanks to J Wielemaker and D Diaz for providing their free Prolog systems Mats Carlsson and Per Mildner supported the development providing help and new capabilities in the ODBC library Also thanks to all the people providing feedback since they are guiding DES to suit more demanded requirements Contributors are specially acknowledged Markus Triska for developing the Emacs IDE and also author of the SWI Prolog clpfd library R Haemmerl for tweaking the Ciao clpfd library and the students Diego Cardiel Freire Juan Jos Ortiz Sanchez Delfin Rup rez Ca as Miguel Martin and Javier Salcedo who developed and improved ACIDE Thanks to Yolanda Garcia and Rafael Caballero for making declarative debugging true for both Datalog and SQL databases They are also key authors in the inclusion of test case generation for SQL views In particular Yolanda took the implementation effort supported by Rafael Gabriel Aranda L pez and Sonia Est vez Martin generated Mac OSX Snow Leopard and Leopard executables resp Enrique Martin Martin fixed the Linux distribution of DES 1 5 0 Finally thanks to the Spanish projects FAST STAMP TIN2008 06622 C03 01 Prometidos CM S2009TIC 1465 and GPD UCM UCM BSCH GR35 10 A 910502 which supported this work Fernando Sdenz P rez 192 204
260. ution p 0 p X p X p 1 The query p X returns the inferred facts from the program irrespective of the apparent infinite recursion in the second rule Note that the Prolog goal p 1 does not terminate You can easily check it out with prolog p 1 6 6 Transitive Closure files tranclosure d1 sql rail With this example we show a possible use of mutual recursion by means of a Datalog program that defines the transitive closure of the relations p and ai It can be consulted with Ze tranclosure p a b p c d q b c q d e pqs X Y p X Y pqs X Y q X Y pqs X Y pqs X Z p Z Y pqs X Y pqs X Z q Z Y The query pqs X Y returns the whole set of inferred facts that model the transitive closure File tranclosure sql contains the SQL counterpart code which can be executed with process tranclosure sql create table p x y insert into p values a b 12 Taken from Diet87 Fernando Sdenz P rez 175 204 Universidad Complutense de Madrid Datalog Educational System insert into p values c d create table q x y insert into q values b c insert into q values d e create view pqs x y as select from p union select from q union select pqs x p y from pqs p where pqs y p x union select pqs x q y from pqs q where pqs y q x The query select from pqs returns the same answer as before File tranclosure ra contains the
261. views or predicates are involved However a given node in the graph will be traversed only once In the case of Datalog queries this graph contains the nodes and edges in the dependency graph restricted to the query ignoring other nodes which do not take part in its computation In the case of SQL the graph shows the dependencies between a view and its data sources in the FROM clause Next tracing for both Datalog queries and SQL views are explained and illustrated with examples 5 7 1 Tracing Datalog Queries The command trace_datalog Goal Order allows to trace a Datalog goal in the given order postorder or the default preorder Goals should be basic i e no conjunctive or disjunctive goals are allowed For instance let s consider the program in the file negation dl and its dependency graph shown in Figure 3 A tracing session could be as follows DES gt c negation Warning Undefined predicate s d 0 Fernando S enz P rez 125 204 Universidad Complutense de Madrid Datalog Educational System DES gt trace_datalog a Info Tracing predicate air a Info 1 tuple in the answer table Info Remaining predicates b 0 c 0 d 0 Input Continue y n yl Info Tracing predicate b not b Info 1 tuple in the answer table Info Remaining predicates c 0 d 0 Input Continue y n yl Info Tracing predicate c c Info 1 tuple in the answer table Info Remaining predicates
262. which is the default database DES deductive engine You can close all connections but the default one As the names suggest you can open a wide range of data sources not only from database management systems as DB2 Oracle SQL Server but also from other sources as datasheets Excel and text files CSV comma separated values files For defining a table in MS Excel you should use Insert gt Name gt Define where you specify the name of the table and the cell range it covers where first row can be used as field names optionally Types are inferred by the Excel system Similarly when defining a connection to a text file field names can be those in the first line of explicitly given Again types are inferred In both cases you can inspect the database schema and query them with either SQL statatements or Datalog queries or RA expressions A warning for newbies You have to define connection names following ODBC installation do not expect the ones listed above are provided by default you need both the ODBC connection and the data provider database server or whatever already installed and configured Fernando Sdenz P rez 98 204 Universidad Complutense de Madrid Datalog Educational System 5 1 4 Current Connection To find out the current opened ODBC database use the command DES SQL gt current_db 5 1 5 Making a Connection the Current One Making a given connection the current one is simply done with D
263. xpected tuple This information is used for slicing the associated queries keeping only those parts that might be the cause of the error The validity of the results produced by sliced queries is easier to determine thus facilitating the location of the error 5 9 2 1 Missing Tuples Let s consider another following example located at examples SQLDebugger examplel sql The loyalty program of an academy awards an intensive course for students that satisfy the following constraints e The student has completed the basic level course level 0 e The student has not completed an intensive course e To complete an intensive course a student must either pass the all in one course or the three initial level courses levels 1 2 and 3 The database schema includes three tables e courses id level contains information about the standard courses including their identifier and the course level e registration student course pass indicates that the student is in the course with pass taking the value true if the course has been successfully completed e allInOneCourse student pass contains information about students registered in a special intensive course with pass playing the same role as in registration File example1 sql contains the SQL views selecting the award candidates The first view is standard which completes the information included in the table registration with the course level The view basic selects those standar
264. y supports test case generation 5 11 Batch Processing There are two ways for processing batch files 1 If the file des ini is located at the distribution directory its contents are interpreted as input prompts and executed before giving control to the user at start up of the system 2 The command process filename or p as a shorthand allows to process each line in the file as it was an input the same way as before If no file extension is given and filename does not exists then ini sql and ra are appended in turn to filename and tried in that order for finding an existing file When processing batch files prompt inputs starting with the symbol are interpreted as comments This way the batch file des ini may contain comments The user can also interactively input such comments but again produce no effects Batch processing can include logging to produce output This is useful to feed the system with batch input and get its output in a file maybe avoiding any interactive input For example consider the following des ini excerpt Dump output to output txt log output txt pretty_print off Process Datalog SQL queries and commands Ze examples fib fib 100 F End log nolog The result found in output txt should be modulo blank lines DES gt pretty_print off Info Pretty print is off DES gt Process Datalog SQL queries and commands DES gt c examples fib Warning N gt 1
265. ysql This retrieves all the data stored in the external database and stores it back in the in memory database of DES In addition to the view p and table p_des_table created in the external database for p there is also a table p_des_metadata holding the Datalog intensional rules that have been made persistent This is needed to recover the original rules as they were asserted in its compiled Datalog form If you have persisted a predicate for which no type constraints has been given before a type constraint is derived if possible and asserted This type constraint remains even when the persistency assertion is removed If you want to remove this too then submit a drop_ic command The following session illustrates this DES gt dbschema Info Database Sdes Info No tables Info No views Info No integrity constraints DES gt persistent p a int mysql DES gt dbschema Info Database Sdes Info No tables Info View s p a number integer Defining SQL statement CREATE VIEW p a AS SELECT ALL FROM p_des_table Info No integrity constraints DES gt drop_assertion persistent p a int mysql DES gt dbschema Fernando Sdenz P rez 110 204 Universidad Complutense de Madrid Datalog Educational System Info Database Sdes Info Table s p a number integer Info No views Info No integrity constraints DES gt drop_ic type p a int DES gt dbschema Info Database Sde

Datalog Educational System V3.0 User`s Manual

Contents

Download Pdf Manuals

Related Search

Related Contents