Package org.apache.druid.segment.column

Interface ColumnCapabilities

All Superinterfaces:

TypeSignature<ValueType>

All Known Implementing Classes:

ColumnCapabilitiesImpl

public interface ColumnCapabilities
extends TypeSignature<ValueType>

This interface is used to expose information about columns that is interesting to know for all matters dealing with reading from columns, including query planning and optimization, creating readers to merge segments at ingestion time, and probably nearly anything else you can imagine.

See Also:

• ColumnType
• TypeSignature

Nested Class Summary

Nested Classes

Modifier and Type	Interface	Description
static enum	ColumnCapabilities.Capable
static interface	ColumnCapabilities.CoercionLogic	This interface defines the shape of a mechanism to allow for bespoke coercion of ColumnCapabilities.Capable.UNKNOWN into ColumnCapabilities.Capable.TRUE or ColumnCapabilities.Capable.FALSE for each ColumnCapabilities.Capable of a ColumnCapabilities, as is appropriate for the situation of the caller.

Method Summary

Modifier and Type	Method	Description
ColumnCapabilities.Capable	areDictionaryValuesSorted()	If the column is dictionary encoded, are those values sorted? Useful to know for optimizations that can defer looking up values and allowing sorting with the dictionary ids directly.
ColumnCapabilities.Capable	areDictionaryValuesUnique()	If the column is dictionary encoded, is there a 1:1 mapping of dictionary ids to values? If this is true, it unlocks optimizations such as allowing for things like grouping directly on dictionary ids and deferred value lookup.
default <T> TypeStrategy<T>	getStrategy()	A TypeStrategy provides facilities to reading and writing values to buffers, as well as basic value comparators and byte size estimation.
boolean	hasBitmapIndexes()	Does the column have an inverted index bitmap for each value? If so, these may be employed to 'pre-filter' the column by examining if the values match the filter and intersecting the bitmaps, to avoid having to scan and evaluate if every row matches the filter.
ColumnCapabilities.Capable	hasMultipleValues()	String columns are sneaky, and might have multiple values, this is to allow callers to know and appropriately prepare themselves.
ColumnCapabilities.Capable	hasNulls()	Does this column contain null values? If so, callers, especially for primitive numeric columns, will need to check for null value rows and act accordingly.
boolean	hasSpatialIndexes()	Does the column have spatial indexes available to allow use with spatial filtering?
ColumnCapabilities.Capable	isDictionaryEncoded()	Is the column dictionary encoded? If so, a DimensionDictionarySelector may be used instead of using a value selector, allowing algorithms to operate on primitive integer dictionary ids rather than the looked up dictionary values.
default ColumnType	toColumnType()

Methods inherited from interface org.apache.druid.segment.column.TypeSignature

anyOf, asTypeString, getComplexTypeName, getElementType, getNullableStrategy, getType, is, isArray, isNumeric, isPrimitive, isPrimitiveArray

Method Details

isDictionaryEncoded

ColumnCapabilities.Capable isDictionaryEncoded()

Is the column dictionary encoded? If so, a DimensionDictionarySelector may be used instead of using a value selector, allowing algorithms to operate on primitive integer dictionary ids rather than the looked up dictionary values.

areDictionaryValuesSorted

ColumnCapabilities.Capable areDictionaryValuesSorted()

If the column is dictionary encoded, are those values sorted? Useful to know for optimizations that can defer looking up values and allowing sorting with the dictionary ids directly.

areDictionaryValuesUnique

ColumnCapabilities.Capable areDictionaryValuesUnique()

If the column is dictionary encoded, is there a 1:1 mapping of dictionary ids to values? If this is true, it unlocks optimizations such as allowing for things like grouping directly on dictionary ids and deferred value lookup.

hasMultipleValues

ColumnCapabilities.Capable hasMultipleValues()

String columns are sneaky, and might have multiple values, this is to allow callers to know and appropriately prepare themselves.

hasBitmapIndexes

boolean hasBitmapIndexes()

Does the column have an inverted index bitmap for each value? If so, these may be employed to 'pre-filter' the column by examining if the values match the filter and intersecting the bitmaps, to avoid having to scan and evaluate if every row matches the filter.

hasSpatialIndexes

boolean hasSpatialIndexes()

Does the column have spatial indexes available to allow use with spatial filtering?

hasNulls

ColumnCapabilities.Capable hasNulls()

Does this column contain null values? If so, callers, especially for primitive numeric columns, will need to check for null value rows and act accordingly.

toColumnType

default ColumnType toColumnType()

getStrategy

default <T> TypeStrategy<T> getStrategy()

Description copied from interface: TypeSignature

A TypeStrategy provides facilities to reading and writing values to buffers, as well as basic value comparators and byte size estimation. Use TypeSignature.getNullableStrategy() if you need to read and write values which might possibly be null and aren't handling this in a different (probably better) way.

Specified by:

getStrategy in interface TypeSignature<ValueType>