wip

Author: hmottestad

core/queryalgebra/evaluation/src/main/java/org/eclipse/rdf4j/query/algebra/evaluation/optimizer/MinusOptimizer.java

@@ -13,7 +13,6 @@
 
 import java.util.ArrayList;
 import java.util.List;
-import java.util.Set;
 
 import org.eclipse.rdf4j.query.BindingSet;
 import org.eclipse.rdf4j.query.Dataset;
@@ -65,10 +64,6 @@ public void meet(Difference difference) {
 				difference.replaceWith(leftArg);
 				return;
 			}
-			if (disjoint(leftArg.getBindingNames(), rightArg.getBindingNames())) {
-				difference.replaceWith(leftArg);
-				return;
-			}
 			if (enableUnionSplit && rightArg instanceof Union) {
 				Union union = (Union) rightArg;
 				List<TupleExpr> branches = new ArrayList<>();
@@ -84,18 +79,6 @@ public void meet(Difference difference) {
 		}
 	}
 
-	private static boolean disjoint(Set<String> leftBindings, Set<String> rightBindings) {
-		if (leftBindings.isEmpty() || rightBindings.isEmpty()) {
-			return true;
-		}
-		for (String name : leftBindings) {
-			if (rightBindings.contains(name)) {
-				return false;
-			}
-		}
-		return true;
-	}
-
 	private static void collectUnionBranches(TupleExpr expr, boolean variableScopeChange, List<TupleExpr> branches) {
 		if (expr instanceof Union && ((Union) expr).isVariableScopeChange() == variableScopeChange) {
 			Union union = (Union) expr;

core/queryalgebra/evaluation/src/main/java/org/eclipse/rdf4j/query/algebra/evaluation/optimizer/sparqluo/BeTreeTransformer.java

@@ -205,6 +205,10 @@ private void applyUnionCommonPrefixPullUp(BeGroupNode group, Set<BeUnionNode> un
 				index++;
 				continue;
 			}
+			if (leavesEmptyBranch(union, common)) {
+				index++;
+				continue;
+			}
 			double baseCost = costEstimator.estimateGroupCost(group);
 			BeUndoToken undo = new BeUndoToken();
 			undo.capture(group);
@@ -394,6 +398,38 @@ private List<StatementPattern> leadingStatementPatterns(BeGroupNode branch) {
 		return patterns;
 	}
 
+	private boolean leavesEmptyBranch(BeUnionNode union, List<StatementPattern> common) {
+		Map<StatementPatternKey, Integer> counts = countPatterns(common);
+		for (BeGroupNode branch : union.getBranches()) {
+			if (branchWouldBeEmpty(branch, new HashMap<>(counts))) {
+				return true;
+			}
+		}
+		return false;
+	}
+
+	private boolean branchWouldBeEmpty(BeGroupNode branch, Map<StatementPatternKey, Integer> removalCounts) {
+		int totalPatterns = 0;
+		for (BeNode child : branch.getChildren()) {
+			if (!(child instanceof BeBgpNode)) {
+				return false;
+			}
+			totalPatterns += ((BeBgpNode) child).getStatementPatterns().size();
+		}
+		int removed = 0;
+		for (BeNode child : branch.getChildren()) {
+			for (StatementPattern pattern : ((BeBgpNode) child).getStatementPatterns()) {
+				StatementPatternKey key = StatementPatternKey.of(pattern);
+				Integer remaining = removalCounts.get(key);
+				if (remaining != null && remaining > 0) {
+					removalCounts.put(key, remaining - 1);
+					removed++;
+				}
+			}
+		}
+		return removed == totalPatterns;
+	}
+
 	private Map<StatementPatternKey, Integer> countPatterns(List<StatementPattern> patterns) {
 		Map<StatementPatternKey, Integer> counts = new HashMap<>();
 		for (StatementPattern pattern : patterns) {

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

@@ -110,7 +110,6 @@ private void assertUnionCommonPrefix(boolean useSparqlUo) {
 		}
 	}
 
-
 	private void assertMinusInsideOptional(boolean useSparqlUo) {
 		SailRepository repo = createRepository(useSparqlUo);
 		try (RepositoryConnection connection = repo.getConnection()) {
@@ -233,7 +232,6 @@ private SailRepository createRepository(boolean useSparqlUo) {
 		return repository;
 	}
 
-
 	private EvaluationStrategyFactory createStandardPipelineFactory(MemoryStore store) {
 		DefaultEvaluationStrategyFactory factory = new DefaultEvaluationStrategyFactory(
 				store.getFederatedServiceResolver()) {