Package org.apache.druid.segment.vector

Interface VectorColumnSelectorFactory

All Superinterfaces:

ColumnInspector, Expr.InputBindingInspector

All Known Implementing Classes:

QueryableIndexVectorColumnSelectorFactory, RemapVectorColumnSelectorFactory

public interface VectorColumnSelectorFactory
extends ColumnInspector

A class that comes from VectorCursor.getColumnSelectorFactory() and is used to create vector selectors. If you need to write code that adapts to different input types, you should write a VectorColumnProcessorFactory and use one of the ColumnProcessors.makeVectorProcessor(java.lang.String, org.apache.druid.segment.VectorColumnProcessorFactory<T>, org.apache.druid.segment.vector.VectorColumnSelectorFactory) functions instead of using this class.

See Also:

• the non-vectorized version.

Method Summary

Modifier and Type	Method	Description
ColumnCapabilities	getColumnCapabilities(String column)	Returns capabilities of a particular column, or null if the column doesn't exist.
default int	getMaxVectorSize()	Returns the maximum vector size for the VectorCursor that generated this object.
ReadableVectorInspector	getReadableVectorInspector()	Returns a ReadableVectorInspector for the VectorCursor that generated this object.
default GroupByVectorColumnSelector	makeGroupByVectorColumnSelector(String column, DeferExpressionDimensions deferExpressionDimensions)	Returns a group-by selector.
MultiValueDimensionVectorSelector	makeMultiValueDimensionSelector(DimensionSpec dimensionSpec)	Returns a dictionary encoded, string-typed, multi-value-per-row column selector.
VectorObjectSelector	makeObjectSelector(String column)	Returns an object selector.
SingleValueDimensionVectorSelector	makeSingleValueDimensionSelector(DimensionSpec dimensionSpec)	Returns a dictionary encoded, string-typed, single-value-per-row column selector.
VectorValueSelector	makeValueSelector(String column)	Returns a primitive column selector.

Methods inherited from interface org.apache.druid.segment.ColumnInspector

getType

Methods inherited from interface org.apache.druid.math.expr.Expr.InputBindingInspector

areNumeric, areNumeric, areSameTypes, areSameTypes, areScalar, areScalar, canVectorize, canVectorize

Method Details

getReadableVectorInspector

ReadableVectorInspector getReadableVectorInspector()

Returns a ReadableVectorInspector for the VectorCursor that generated this object.

getMaxVectorSize

default int getMaxVectorSize()

Returns the maximum vector size for the VectorCursor that generated this object.

See Also:

• VectorSizeInspector.getMaxVectorSize()

makeSingleValueDimensionSelector

SingleValueDimensionVectorSelector makeSingleValueDimensionSelector(DimensionSpec dimensionSpec)

Returns a dictionary encoded, string-typed, single-value-per-row column selector. Should only be called on columns where getColumnCapabilities(java.lang.String) indicates they return STRING, or on nonexistent columns. Since the selector created with this method operates directly on the dictionary encoded input, STRING values must be translated from the dictionary id for a given row value using DimensionDictionarySelector.lookupName(int), but this selector can prove optimal for operations which can be done directly on the underlying dictionary ids, such as grouping within a segment. If you need to write code that adapts to different input types, you should write a VectorColumnProcessorFactory and use one of the ColumnProcessors.makeVectorProcessor(java.lang.String, org.apache.druid.segment.VectorColumnProcessorFactory<T>, org.apache.druid.segment.vector.VectorColumnSelectorFactory) functions instead of using this method.

makeMultiValueDimensionSelector

MultiValueDimensionVectorSelector makeMultiValueDimensionSelector(DimensionSpec dimensionSpec)

Returns a dictionary encoded, string-typed, multi-value-per-row column selector. Should only be called on columns where getColumnCapabilities(java.lang.String) indicates they return STRING. Unlike makeSingleValueDimensionSelector(org.apache.druid.query.dimension.DimensionSpec), this should not be called on nonexistent columns. If you need to write code that adapts to different input types, you should write a VectorColumnProcessorFactory and use one of the ColumnProcessors.makeVectorProcessor(java.lang.String, org.apache.druid.segment.VectorColumnProcessorFactory<T>, org.apache.druid.segment.vector.VectorColumnSelectorFactory) functions instead of using this method.

makeValueSelector

VectorValueSelector makeValueSelector(String column)

Returns a primitive column selector. Should only be called on columns where getColumnCapabilities(java.lang.String) indicates they return DOUBLE, FLOAT, or LONG, or on nonexistent columns. If you need to write code that adapts to different input types, you should write a VectorColumnProcessorFactory and use one of the ColumnProcessors.makeVectorProcessor(java.lang.String, org.apache.druid.segment.VectorColumnProcessorFactory<T>, org.apache.druid.segment.vector.VectorColumnSelectorFactory) functions instead of using this method.

makeObjectSelector

VectorObjectSelector makeObjectSelector(String column)

Returns an object selector. Should only be called on columns where getColumnCapabilities(java.lang.String) indicates that they return STRING, ARRAY, or COMPLEX, or on nonexistent columns. For STRING, this is needed if values are not dictionary encoded, such as computed virtual columns, or can optionally be used in place of SingleValueDimensionVectorSelector when using the dictionary isn't helpful. Currently, this should only be called on single valued STRING inputs (multi-value STRING vector object selector is not yet implemented). If you need to write code that adapts to different input types, you should write a VectorColumnProcessorFactory and use one of the ColumnProcessors.makeVectorProcessor(java.lang.String, org.apache.druid.segment.VectorColumnProcessorFactory<T>, org.apache.druid.segment.vector.VectorColumnSelectorFactory) functions instead of using this method.

getColumnCapabilities

@Nullable
ColumnCapabilities getColumnCapabilities(String column)

Returns capabilities of a particular column, or null if the column doesn't exist. Unlike ColumnSelectorFactory, null does not potentially indicate a dynamically discovered column.

Specified by:

getColumnCapabilities in interface ColumnInspector

Parameters:

column - column name

Returns:

capabilities, or null if the column doesn't exist.

makeGroupByVectorColumnSelector

default GroupByVectorColumnSelector makeGroupByVectorColumnSelector(String column,
 DeferExpressionDimensions deferExpressionDimensions)

Returns a group-by selector. Allows columns to control their own grouping behavior.

Parameters:

column - column name

deferExpressionDimensions - active value of GroupByQueryConfig.CTX_KEY_DEFER_EXPRESSION_DIMENSIONS