cweb-cvs

[Bryan T. <tho...@us...>]

Update of /cvsroot/cweb/bigdata-rdf/src/java/com/bigdata/rdf/inf
In directory sc8-pr-cvs4.sourceforge.net:/tmp/cvs-serv3033/src/java/com/bigdata/rdf/inf

Modified Files:
	RuleRdfs02.java AbstractRuleRdf.java Rule.java RuleRdfs09.java 
	RuleRdf01.java RuleRdfs03.java 
Log Message:
Optimization for Rule rdf1.

Index: RuleRdf01.java
===================================================================
RCS file: /cvsroot/cweb/bigdata-rdf/src/java/com/bigdata/rdf/inf/RuleRdf01.java,v
retrieving revision 1.9
retrieving revision 1.10
diff -C2 -d -r1.9 -r1.10
*** RuleRdf01.java	18 Apr 2007 17:29:07 -0000	1.9
--- RuleRdf01.java	19 Apr 2007 19:13:45 -0000	1.10
***************
*** 44,48 ****
  package com.bigdata.rdf.inf;
  
! import com.bigdata.btree.IEntryIterator;
  import com.bigdata.rdf.KeyOrder;
  
--- 44,50 ----
  package com.bigdata.rdf.inf;
  
! import java.util.ArrayList;
! import java.util.Iterator;
! 
  import com.bigdata.rdf.KeyOrder;
  
***************
*** 70,111 ****
          final long computeStart = System.currentTimeMillis();
          
!         long lastP = -1;
!         
!         /*
!          * This is essentially doing a "select distinct predicate".
!          * 
!          * FIXME there should be a lighter weight way of achieving this result.
!          * One way would be to create a "predicates" index that had each
!          * distinct predicate (further denormalizing the schema and requiring
!          * maintenance). Another approach is to restart the iterator each time a
!          * predicate [p] is found by computing the fromKey as [p+1] thereby
!          * skipping over all intervening statements in the index (consider how
!          * to distribute that query in parallel).
!          * 
!          * @todo write a test for this rule and then write an optimized variant
!          * using the incremented toKey approach.
!          */
!         IEntryIterator it = store.getPOSIndex().rangeIterator(null,null); 
!         
!         while ( it.hasNext() ) {
!             
!             it.next();
!             
!             stats.stmts1++;
              
!             SPO stmt = new SPO(KeyOrder.POS, store.keyBuilder, it.getKey());
              
!             if ( stmt.p != lastP ) {
!                 
!                 lastP = stmt.p;
  
-                 buffer.add(new SPO(stmt.p, store.rdfType.id,
-                         store.rdfProperty.id));
-                 
                  stats.numComputed++;
!                 
              }
!             
!         }
          
          stats.computeTime += System.currentTimeMillis() - computeStart;
--- 72,135 ----
          final long computeStart = System.currentTimeMillis();
          
! //        if (false) {
! //            
! //            /*
! //             * Original implementation does a full scan of all statements in the
! //             * KB.
! //             */
! //            
! //            long lastP = NULL;
! //
! //            IEntryIterator it = store.getPOSIndex().rangeIterator(null, null);
! //
! //            while (it.hasNext()) {
! //
! //                it.next();
! //
! //                stats.stmts1++;
! //
! //                SPO stmt = new SPO(KeyOrder.POS, store.keyBuilder, it.getKey());
! //
! //                if (stmt.p != lastP) {
! //
! //                    lastP = stmt.p;
! //
! //                    stats.numComputed++;
! //
! //                    System.err.println("" + stats.numComputed + " : " + lastP
! //                            + " : " + store.toString(lastP));
! //
! //                    buffer.add(new SPO(stmt.p, store.rdfType.id,
! //                            store.rdfProperty.id));
! //
! //                }
! //
! //            }
! //            
! //        } else {
              
!             /*
!              * Alternative implementation does an efficient scan for only the
!              * distinct predicates in the KB.
!              */
              
!             // find the distinct predicates in the KB.
!             ArrayList<Long> ids = store.distinctTermScan(KeyOrder.POS);
! 
!             Iterator<Long> itr = ids.iterator();
! 
!             while (itr.hasNext()) {
! 
!                 stats.stmts1++;
! 
!                 long p = itr.next();
! 
!                 buffer.add(new SPO(p, store.rdfType.id, store.rdfProperty.id));
  
                  stats.numComputed++;
! 
              }
! 
! //        }
          
          stats.computeTime += System.currentTimeMillis() - computeStart;

Index: RuleRdfs02.java
===================================================================
RCS file: /cvsroot/cweb/bigdata-rdf/src/java/com/bigdata/rdf/inf/RuleRdfs02.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** RuleRdfs02.java	18 Apr 2007 17:29:07 -0000	1.2
--- RuleRdfs02.java	19 Apr 2007 19:13:45 -0000	1.3
***************
*** 44,49 ****
  package com.bigdata.rdf.inf;
  
! 
! 
  public class RuleRdfs02 extends AbstractRuleRdfs_2_3_7_9 {
  
--- 44,57 ----
  package com.bigdata.rdf.inf;
  
! /**
!  * rdfs2:
!  * <pre>
!  * triple( u rdf:type x) :-
!  *    triple( a rdfs:domain x),
!  *    triple( u a y ).
!  * </pre>
!  * @author <a href="mailto:tho...@us...">Bryan Thompson</a>
!  * @version $Id$
!  */
  public class RuleRdfs02 extends AbstractRuleRdfs_2_3_7_9 {
  

Index: RuleRdfs09.java
===================================================================
RCS file: /cvsroot/cweb/bigdata-rdf/src/java/com/bigdata/rdf/inf/RuleRdfs09.java,v
retrieving revision 1.5
retrieving revision 1.6
diff -C2 -d -r1.5 -r1.6
*** RuleRdfs09.java	18 Apr 2007 17:29:07 -0000	1.5
--- RuleRdfs09.java	19 Apr 2007 19:13:45 -0000	1.6
***************
*** 53,59 ****
   *          triple(?v,rdf:type,?u). 
   * </pre>
-  * <pre>
-  *  &lt;u rdfs:subClassOf x&gt; AND &lt;v rdf:type u&gt; IMPLIES &lt;v rdf:type x&gt;
-  * </pre>
   */
  public class RuleRdfs09 extends AbstractRuleRdfs_2_3_7_9 {
--- 53,56 ----

Index: Rule.java
===================================================================
RCS file: /cvsroot/cweb/bigdata-rdf/src/java/com/bigdata/rdf/inf/Rule.java,v
retrieving revision 1.7
retrieving revision 1.8
diff -C2 -d -r1.7 -r1.8
*** Rule.java	18 Apr 2007 17:29:07 -0000	1.7
--- Rule.java	19 Apr 2007 19:13:45 -0000	1.8
***************
*** 178,181 ****
--- 178,188 ----
          long computeTime;
          
+         public String toString() {
+             
+             return ", #stmts1=" + stmts1 + ", #stmts2=" + stmts2
+                     + ", #subqueries=" + numSubqueries + ", numConsidered+"
+                     + numConsidered + ", computeTime=" + computeTime;
+         }
+ 
      }
  

Index: AbstractRuleRdf.java
===================================================================
RCS file: /cvsroot/cweb/bigdata-rdf/src/java/com/bigdata/rdf/inf/AbstractRuleRdf.java,v
retrieving revision 1.8
retrieving revision 1.9
diff -C2 -d -r1.8 -r1.9
*** AbstractRuleRdf.java	14 Apr 2007 13:33:51 -0000	1.8
--- AbstractRuleRdf.java	19 Apr 2007 19:13:45 -0000	1.9
***************
*** 44,51 ****
  package com.bigdata.rdf.inf;
  
- import org.openrdf.model.URI;
- 
- import com.bigdata.btree.IIndex;
- 
  public abstract class AbstractRuleRdf extends Rule {
      
--- 44,47 ----

Index: RuleRdfs03.java
===================================================================
RCS file: /cvsroot/cweb/bigdata-rdf/src/java/com/bigdata/rdf/inf/RuleRdfs03.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** RuleRdfs03.java	18 Apr 2007 17:29:07 -0000	1.2
--- RuleRdfs03.java	19 Apr 2007 19:13:45 -0000	1.3
***************
*** 44,49 ****
  package com.bigdata.rdf.inf;
  
! 
! 
  public class RuleRdfs03 extends AbstractRuleRdfs_2_3_7_9 {
  
--- 44,59 ----
  package com.bigdata.rdf.inf;
  
! /**
!  * rdfs3:
!  * 
!  * <pre>
!  * triple(v rdf:type x) :-
!  *    triple(a rdfs:range x),
!  *    triple(u a v).
!  * </pre>
!  * 
!  * @author <a href="mailto:tho...@us...">Bryan Thompson</a>
!  * @version $Id$
!  */
  public class RuleRdfs03 extends AbstractRuleRdfs_2_3_7_9 {