com.bigdata.rdf.sparql.ast.optimizers

Interface ASTStaticJoinOptimizer.Annotations

All Superinterfaces:

AST2BOpBase.Annotations

Enclosing class:

ASTStaticJoinOptimizer

public static interface ASTStaticJoinOptimizer.Annotations
extends AST2BOpBase.Annotations

Field Summary

Fields

Modifier and Type	Field and Description
static Double	DEFAULT_OPTIMISTIC See OPTIMISTIC.
static String	OPTIMISTIC The value of this query hint determines how optimistic the optimizer will be in selecting the join cardinality for its joins.

Fields inherited from interface com.bigdata.rdf.sparql.ast.eval.AST2BOpBase.Annotations

AUTO_COMMIT, COST_SCAN, COST_SUBQUERY, DATASET, DEFAULT_QUADS, ESTIMATED_CARDINALITY, NKNOWN, ORIGINAL_INDEX, QUADS, SCOPE, SIMPLE_JOIN

Field Detail

OPTIMISTIC

static final String OPTIMISTIC

The value of this query hint determines how optimistic the optimizer will be in selecting the join cardinality for its joins. Basically when there is a join that has both shared and unshared variables, the join cardinality will be somewhere in between the cardinality of the two sides (range count for a statement pattern, previous join cardinality for a predecessor join). The default value is 1.0, which is the historical behavior.

For a value of 0.67 the optimizer takes a mostly optimistic view by default - the join cardinality will be 0.67 * the MIN + 0.33 * the MAX. This settting will eliminate some of the worst possible outcomes (ones where we guess wrong and get a very bad join order as a result).

BSBM BI Q1 is a good example of a query that benefits from the pessimistic approach, and LUBM Q2 is a good example of a query that benefits from the optimistic approach.

DEFAULT_OPTIMISTIC

static final Double DEFAULT_OPTIMISTIC

See OPTIMISTIC.