GH-2140 Unions can now be created without a new scope and will be han…

…dled more efficiently.

Signed-off-by: Håvard Ottestad <hmottestad@gmail.com>

core/queryalgebra/evaluation/src/main/java/org/eclipse/rdf4j/query/algebra/evaluation/impl/QueryModelNormalizer.java

@@ -36,7 +36,7 @@
 /**
  * A query optimizer that (partially) normalizes query models to a canonical form. Note: this implementation does not
  * yet cover all query node types.
- * 
+ *
  * @author Arjohn Kampman
  */
 public class QueryModelNormalizer extends AbstractQueryModelVisitor<RuntimeException> implements QueryOptimizer {
@@ -67,15 +67,15 @@ public void meet(Join join) {
  Union union = (Union) leftArg;
  Join leftJoin = new Join(union.getLeftArg(), rightArg.clone());
  Join rightJoin = new Join(union.getRightArg(), rightArg.clone());
- Union newUnion = new Union(leftJoin, rightJoin);
+ Union newUnion = new Union(leftJoin, rightJoin, union.isNewScope());
  join.replaceWith(newUnion);
  newUnion.visit(this);
  } else if (rightArg instanceof Union) {
  // sort unions above joins
  Union union = (Union) rightArg;
  Join leftJoin = new Join(leftArg.clone(), union.getLeftArg());
  Join rightJoin = new Join(leftArg.clone(), union.getRightArg());
- Union newUnion = new Union(leftJoin, rightJoin);
+ Union newUnion = new Union(leftJoin, rightJoin, union.isNewScope());
  join.replaceWith(newUnion);
  newUnion.visit(this);
  } else if (leftArg instanceof LeftJoin && isWellDesigned(((LeftJoin) leftArg))) {

core/queryalgebra/evaluation/src/main/java/org/eclipse/rdf4j/query/algebra/evaluation/iterator/JoinIterator.java

@@ -105,7 +105,7 @@ protected void handleClose() throws QueryEvaluationException {
  }
 
  private boolean isOutOfScopeForLeftArgBindings(TupleExpr expr) {
- if (expr instanceof Union) {
+ if (expr instanceof Union && ((Union) expr).isNewScope()) {
  return true;
  }
  return TupleExprs.isGraphPatternGroup(expr) && !(expr instanceof Filter);

core/queryalgebra/model/src/main/java/org/eclipse/rdf4j/query/algebra/Union.java

@@ -15,21 +15,36 @@
  */
 public class Union extends BinaryTupleOperator {
 
+ // Used to differentiate between Unions introduced by the user that require a new scope and unions introduced for
+ // evaluation or performance purposes that do not introduce a new scope. This means that they can be executed more
+ // efficiently.
+ private boolean newScope;
+
  /*--------------*
  * Constructors *
  *--------------*/
 
  public Union() {
+ this(true);
+ }
+
+ public Union(boolean newScope) {
+ this.newScope = newScope;
  }
 
  /**
  * Creates a new union operator that operates on the two specified arguments.
- * 
+ *
  * @param leftArg The left argument of the union operator.
  * @param rightArg The right argument of the union operator.
  */
  public Union(TupleExpr leftArg, TupleExpr rightArg) {
+ this(leftArg, rightArg, true);
+ }
+
+ public Union(TupleExpr leftArg, TupleExpr rightArg, boolean newScope) {
  super(leftArg, rightArg);
+ this.newScope = newScope;
  }
 
  /*---------*
@@ -69,6 +84,16 @@ public int hashCode() {
 
  @Override
  public Union clone() {
- return (Union) super.clone();
+ Union clone = (Union) super.clone();
+ clone.newScope = this.newScope;
+ return clone;
+ }
+
+ public boolean isNewScope() {
+ return newScope;
+ }
+
+ public void setNewScope(boolean newScope) {
+ this.newScope = newScope;
  }
 }

core/queryparser/sparql/src/main/java/org/eclipse/rdf4j/query/parser/sparql/TupleExprBuilder.java

@@ -1187,7 +1187,7 @@ public Object visit(ASTPathAlternative pathAltNode, Object data) throws VisitorE
 
  if (altCount > 1) {
  GraphPattern parentGP = graphPattern;
- Union union = new Union();
+ Union union = new Union(false);
  Union currentUnion = union;
  for (int i = 0; i < altCount - 1; i++) {
  graphPattern = new GraphPattern(parentGP);
@@ -1200,7 +1200,7 @@ public Object visit(ASTPathAlternative pathAltNode, Object data) throws VisitorE
  arg = graphPattern.buildTupleExpr();
  currentUnion.setRightArg(arg);
  } else {
- Union newUnion = new Union();
+ Union newUnion = new Union(false);
  currentUnion.setRightArg(newUnion);
  currentUnion = newUnion;
  }
@@ -1595,7 +1595,7 @@ private TupleExpr handlePathModifiers(Scope scope, Var subjVar, TupleExpr te, Va
  // upperbound is fixed-length
 
  // create set of unions for all path lengths between lower and upper bound.
- Union union = new Union();
+ Union union = new Union(false);
  Union currentUnion = union;
 
  for (long length = lowerBound; length < upperBound; length++) {
@@ -1607,7 +1607,7 @@ private TupleExpr handlePathModifiers(Scope scope, Var subjVar, TupleExpr te, Va
  path = createPath(scope, subjVar, te, endVar, contextVar, length + 1);
  currentUnion.setRightArg(path);
  } else {
- Union nextUnion = new Union();
+ Union nextUnion = new Union(false);
  currentUnion.setRightArg(nextUnion);
  currentUnion = nextUnion;
  }

core/sail/memory/src/test/java/org/eclipse/rdf4j/sail/memory/QueryPlanRetrievalTest.java

@@ -18,7 +18,6 @@
 
 import org.apache.commons.io.IOUtils;
 import org.eclipse.rdf4j.IsolationLevels;
-import org.eclipse.rdf4j.common.concurrent.locks.Properties;
 import org.eclipse.rdf4j.model.ValueFactory;
 import org.eclipse.rdf4j.model.impl.SimpleValueFactory;
 import org.eclipse.rdf4j.model.vocabulary.FOAF;
@@ -619,4 +618,94 @@ public void testTimeout() {
 
  }
 
+ @Test
+ public void testAlternatePath() {
+ SailRepository sailRepository = new SailRepository(new MemoryStore());
+ try (SailRepositoryConnection connection = sailRepository.getConnection()) {
+ connection.begin();
+ for (int i = 0; i < 10; i++) {
+ connection.add(vf.createBNode(i + ""), RDF.TYPE, vf.createBNode((i + 1) + ""));
+ connection.add(vf.createBNode(i + ""), RDF.TYPE, vf.createBNode((i - 1) + ""));
+
+ connection.add(vf.createBNode(i + ""), RDF.TYPE, vf.createBNode((i + 2) + ""));
+ connection.add(vf.createBNode(i + ""), RDF.TYPE, vf.createBNode((i - 2) + ""));
+ }
+
+ connection.add(vf.createBNode("0"), RDFS.LABEL, vf.createLiteral("label"));
+
+ connection.commit();
+ }
+
+ try (SailRepositoryConnection connection = sailRepository.getConnection()) {
+ Query query = connection.prepareTupleQuery(String.join("\n", "",
+ "PREFIX rdfs:<http://www.w3.org/2000/01/rdf-schema#>",
+ "select * where {",
+ " ?a (a|^a)* ?c. ?c (a|^a) ?b .",
+ " ?b rdfs:label ?type .",
+ "}"));
+
+ query.setMaxExecutionTime(1);
+
+ String actual = query.explain(Explanation.Level.Executed).toString();
+
+ assertEquals("Projection (resultSizeActual=84)\n" +
+ " ProjectionElemList\n" +
+ " ProjectionElem \"a\"\n" +
+ " ProjectionElem \"c\"\n" +
+ " ProjectionElem \"b\"\n" +
+ " ProjectionElem \"type\"\n" +
+ " Join (JoinIterator) (resultSizeActual=84)\n" +
+ " StatementPattern (costEstimate=1, resultSizeEstimate=1, resultSizeActual=1)\n" +
+ " Var (name=b)\n" +
+ " Var (name=_const_9285ccfc_uri, value=http://www.w3.org/2000/01/rdf-schema#label, anonymous)\n"
+ +
+ " Var (name=type)\n" +
+ " Union (resultSizeActual=84)\n" +
+ " Join (JoinIterator) (resultSizeActual=28)\n" +
+ " StatementPattern (costEstimate=2, resultSizeEstimate=40, resultSizeActual=2)\n" +
+ " Var (name=c)\n" +
+ " Var (name=_const_f5e5585a_uri, value=http://www.w3.org/1999/02/22-rdf-syntax-ns#type, anonymous)\n"
+ +
+ " Var (name=b)\n" +
+ " ArbitraryLengthPath (costEstimate=9, resultSizeEstimate=82, resultSizeActual=28)\n" +
+ " Var (name=a)\n" +
+ " Union\n" +
+ " StatementPattern (resultSizeEstimate=40)\n" +
+ " Var (name=a)\n" +
+ " Var (name=_const_f5e5585a_uri, value=http://www.w3.org/1999/02/22-rdf-syntax-ns#type, anonymous)\n"
+ +
+ " Var (name=c)\n" +
+ " StatementPattern (resultSizeEstimate=40)\n" +
+ " Var (name=c)\n" +
+ " Var (name=_const_f5e5585a_uri, value=http://www.w3.org/1999/02/22-rdf-syntax-ns#type, anonymous)\n"
+ +
+ " Var (name=a)\n" +
+ " Var (name=c)\n" +
+ " Join (JoinIterator) (resultSizeActual=56)\n" +
+ " StatementPattern (costEstimate=2, resultSizeEstimate=40, resultSizeActual=4)\n" +
+ " Var (name=b)\n" +
+ " Var (name=_const_f5e5585a_uri, value=http://www.w3.org/1999/02/22-rdf-syntax-ns#type, anonymous)\n"
+ +
+ " Var (name=c)\n" +
+ " ArbitraryLengthPath (costEstimate=9, resultSizeEstimate=82, resultSizeActual=56)\n" +
+ " Var (name=a)\n" +
+ " Union\n" +
+ " StatementPattern (resultSizeEstimate=40)\n" +
+ " Var (name=a)\n" +
+ " Var (name=_const_f5e5585a_uri, value=http://www.w3.org/1999/02/22-rdf-syntax-ns#type, anonymous)\n"
+ +
+ " Var (name=c)\n" +
+ " StatementPattern (resultSizeEstimate=40)\n" +
+ " Var (name=c)\n" +
+ " Var (name=_const_f5e5585a_uri, value=http://www.w3.org/1999/02/22-rdf-syntax-ns#type, anonymous)\n"
+ +
+ " Var (name=a)\n" +
+ " Var (name=c)\n", actual);
+
+ System.out.println(actual);
+ }
+ sailRepository.shutDown();
+
+ }
+
 }