Package org.apache.drill.exec.planner.index

Class IndexConditionInfo.Builder

java.lang.Object

org.apache.drill.exec.planner.index.IndexConditionInfo.Builder

Enclosing class:

IndexConditionInfo

public static class IndexConditionInfo.Builder
extends Object

Constructor Summary

Constructors

Constructor	Description
Builder(org.apache.calcite.rex.RexNode condition, Iterable<IndexDescriptor> indexes, org.apache.calcite.rex.RexBuilder builder, org.apache.calcite.rel.RelNode scan)
Builder(org.apache.calcite.rex.RexNode condition, IndexDescriptor index, org.apache.calcite.rex.RexBuilder builder, DrillScanRel scan)

Method Summary

Modifier and Type	Method	Description
IndexConditionInfo	getCollectiveInfo(IndexLogicalPlanCallContext indexContext)	Get a single IndexConditionInfo in which indexCondition has field on all indexes in this.indexes
Map<IndexDescriptor,IndexConditionInfo>	getFirstKeyIndexConditionMap()	Get a map of Index=>IndexConditionInfo, each IndexConditionInfo has the separated condition and remainder condition.
Map<IndexDescriptor,IndexConditionInfo>	getIndexConditionMap()	Get a map of Index=>IndexConditionInfo, each IndexConditionInfo has the separated condition and remainder condition.
Map<IndexDescriptor,IndexConditionInfo>	getIndexConditionMap(List<IndexDescriptor> indexList)	Get a map of Index=>IndexConditionInfo, each IndexConditionInfo has the separated condition and remainder condition.
IndexConditionInfo	indexConditionRelatedToFields(List<LogicalExpression> relevantPaths, org.apache.calcite.rex.RexNode condition)	Given a list of Index Expressions(usually indexed fields/functions from one or a set of indexes), separate a filter condition into 1), relevant subset of conditions (by relevant, it means at least one given index Expression was found) and, 2), the rest in remainderCondition
boolean	isConditionPrefix(IndexDescriptor indexDesc, org.apache.calcite.rex.RexNode initCondition)	Given a RexNode corresponding to the condition expression tree and the index descriptor, check if one or more columns involved in the condition tree form a prefix of the columns in the index keys.
boolean	isValidIndexHint(IndexLogicalPlanCallContext indexContext)

Methods inherited from class java.lang.Object

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

Constructor Details

Builder

public Builder(org.apache.calcite.rex.RexNode condition,
 Iterable<IndexDescriptor> indexes,
 org.apache.calcite.rex.RexBuilder builder,
 org.apache.calcite.rel.RelNode scan)

Builder

public Builder(org.apache.calcite.rex.RexNode condition,
 IndexDescriptor index,
 org.apache.calcite.rex.RexBuilder builder,
 DrillScanRel scan)

Method Details

getCollectiveInfo

public IndexConditionInfo getCollectiveInfo(IndexLogicalPlanCallContext indexContext)

Get a single IndexConditionInfo in which indexCondition has field on all indexes in this.indexes

Returns:

isValidIndexHint

public boolean isValidIndexHint(IndexLogicalPlanCallContext indexContext)

getFirstKeyIndexConditionMap

public Map<IndexDescriptor,IndexConditionInfo> getFirstKeyIndexConditionMap()

Get a map of Index=>IndexConditionInfo, each IndexConditionInfo has the separated condition and remainder condition. The map is ordered, so the last IndexDescriptor will have the final remainderCondition after separating conditions that are relevant to this.indexes. The conditions are separated on LEADING index columns.

Returns:

Map containing indexIndexDescriptor and condition IndexConditionInfo pairs

isConditionPrefix

public boolean isConditionPrefix(IndexDescriptor indexDesc,
 org.apache.calcite.rex.RexNode initCondition)

Given a RexNode corresponding to the condition expression tree and the index descriptor, check if one or more columns involved in the condition tree form a prefix of the columns in the index keys.

Parameters:

indexDesc -

initCondition -

Returns:

True if prefix, False if not

getIndexConditionMap

public Map<IndexDescriptor,IndexConditionInfo> getIndexConditionMap(List<IndexDescriptor> indexList)

Get a map of Index=>IndexConditionInfo, each IndexConditionInfo has the separated condition and remainder condition. The map is ordered, so the last IndexDescriptor will have the final remainderCondition after separating conditions that are relevant to the indexList. The conditions are separated based on index columns.

Returns:

Map containing indexIndexDescriptor and condition IndexConditionInfo pairs

getIndexConditionMap

public Map<IndexDescriptor,IndexConditionInfo> getIndexConditionMap()

Get a map of Index=>IndexConditionInfo, each IndexConditionInfo has the separated condition and remainder condition. The map is ordered, so the last IndexDescriptor will have the final remainderCondition after separating conditions that are relevant to this.indexes. The conditions are separated based on index columns.

Returns:

Map containing indexIndexDescriptor and condition IndexConditionInfo pairs

indexConditionRelatedToFields

public IndexConditionInfo indexConditionRelatedToFields(List<LogicalExpression> relevantPaths,
 org.apache.calcite.rex.RexNode condition)

Given a list of Index Expressions(usually indexed fields/functions from one or a set of indexes), separate a filter condition into 1), relevant subset of conditions (by relevant, it means at least one given index Expression was found) and, 2), the rest in remainderCondition

Parameters:

relevantPaths -

condition -

Returns: