Package org.eclipse.rdf4j.query.algebra.evaluation.optimizer

Class QueryJoinOptimizer

java.lang.Object

org.eclipse.rdf4j.query.algebra.evaluation.optimizer.QueryJoinOptimizer

All Implemented Interfaces:

QueryOptimizer

public class QueryJoinOptimizer
extends Object
implements QueryOptimizer

A query optimizer that re-orders nested Joins.

Author:

Arjohn Kampman, James Leigh

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
protected class	QueryJoinOptimizer.JoinVisitor	This can be extended by subclasses to allow for adjustments to the optimization process.

Field Summary

Fields

Modifier and Type	Field	Description
static int	MERGE_JOIN_CARDINALITY_SIZE_DIFF_MULTIPLIER	When deciding if merge join is the correct approach we will compare the cardinality of the two join arguments, if one is bigger than the other by a factor of MERGE_JOIN_CARDINALITY_SIZE_DIFF_MULTIPLIER then we will not use merge join.
protected final EvaluationStatistics	statistics
static boolean	USE_MERGE_JOIN_FOR_LAST_STATEMENT_PATTERNS_WHEN_CROSS_JOIN

Constructor Summary

Constructors

Constructor	Description
QueryJoinOptimizer(EvaluationStatistics statistics)
QueryJoinOptimizer(EvaluationStatistics statistics, boolean trackResultSize)
QueryJoinOptimizer(EvaluationStatistics statistics, boolean trackResultSize, TripleSource tripleSource)
QueryJoinOptimizer(EvaluationStatistics statistics, TripleSource tripleSource)

Method Summary

Modifier and Type	Method	Description
void	optimize(TupleExpr tupleExpr, Dataset dataset, BindingSet bindings)	Applies generally applicable optimizations: path expressions are sorted from more to less specific.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Details

MERGE_JOIN_CARDINALITY_SIZE_DIFF_MULTIPLIER

@Experimental
public static int MERGE_JOIN_CARDINALITY_SIZE_DIFF_MULTIPLIER

When deciding if merge join is the correct approach we will compare the cardinality of the two join arguments, if one is bigger than the other by a factor of MERGE_JOIN_CARDINALITY_SIZE_DIFF_MULTIPLIER then we will not use merge join. As an example, if the limit is 10 and the left cardinality if 50 000 and the right cardinality is 500 000 then we will use merge join, but if it is 500 001 then we will not.

USE_MERGE_JOIN_FOR_LAST_STATEMENT_PATTERNS_WHEN_CROSS_JOIN

@Experimental
public static boolean USE_MERGE_JOIN_FOR_LAST_STATEMENT_PATTERNS_WHEN_CROSS_JOIN

statistics

protected final EvaluationStatistics statistics

Constructor Details

QueryJoinOptimizer

public QueryJoinOptimizer(EvaluationStatistics statistics)

QueryJoinOptimizer

public QueryJoinOptimizer(EvaluationStatistics statistics,
 TripleSource tripleSource)

QueryJoinOptimizer

public QueryJoinOptimizer(EvaluationStatistics statistics,
 boolean trackResultSize)

QueryJoinOptimizer

public QueryJoinOptimizer(EvaluationStatistics statistics,
 boolean trackResultSize,
 TripleSource tripleSource)

Method Details

optimize

public void optimize(TupleExpr tupleExpr,
 Dataset dataset,
 BindingSet bindings)

Applies generally applicable optimizations: path expressions are sorted from more to less specific.

Specified by:

optimize in interface QueryOptimizer

Parameters:

tupleExpr -