public class AggregateFilterTransposeRule extends RelOptRule
Filterand transposes them, pushing the aggregate below the filter.
In some cases, it is necessary to split the aggregate.
This rule does not directly improve performance. The aggregate will have to process more rows, to produce aggregated rows that will be thrown away. The rule might be beneficial if the predicate is very expensive to evaluate. The main use of the rule is to match a query that has a filter under an aggregate to an existing aggregate table.
|Modifier and Type||Field and Description|
|Constructor and Description|
Creates an AggregateFilterTransposeRule.
|Modifier and Type||Method and Description|
Receives notification about a rule match.
any, convert, convert, convertList, convertOperand, equals, equals, getOperand, getOperands, getOutConvention, getOutTrait, hashCode, matches, none, operand, operand, operand, operand, operand, some, toString, unordered
public static final AggregateFilterTransposeRule INSTANCE
public void onMatch(RelOptRuleCall call)
call.relsholds the set of relational expressions which match the operands to the rule;
call.relsis the root expression.
Typically a rule would check that the nodes are valid matches, creates
a new expression, then calls back
RelOptRuleCall.transformTo(org.apache.calcite.rel.RelNode, java.util.Map<org.apache.calcite.rel.RelNode, org.apache.calcite.rel.RelNode>) to
register the expression.
Copyright © 2012–2018 The Apache Software Foundation. All rights reserved.