Module org.elasticsearch.compute

Package org.elasticsearch.compute.aggregation.blockhash

Class LongBytesRefAdaptiveBlockHash

java.lang.Object

org.elasticsearch.compute.aggregation.blockhash.BlockHash

org.elasticsearch.compute.aggregation.blockhash.LongBytesRefAdaptiveBlockHash

All Implemented Interfaces:

Closeable, AutoCloseable, SeenGroupIds, Releasable

public final class LongBytesRefAdaptiveBlockHash
extends BlockHash

An AdaptiveBlockHash for the two-key (LONG, BYTES_REF) (or (BYTES_REF, LONG)) grouping shape. As long as both inputs arrive as vectors, it routes through a dictionary-aware fast path: BytesRefBlockHash hashes the bytes column once (so for an OrdinalBytesRefVector it pays per dictionary entry, not per row), and a LongLongHashTable combines (bytesRefOrd, longKey) into a stable group id. As soon as a page brings a non-vector key column (i.e. nulls or multivalues are possible), it migrates to a PackedValuesBlockHash — which has correct null/MV semantics — and replays existing groups so ordinals stay stable.

Modeled after LongIntAdaptiveBlockHash. The migration encoding is variable-width because the BYTES_REF column has no fixed size; everything else is the same.

Nested Class Summary

Nested classes/interfaces inherited from class org.elasticsearch.compute.aggregation.blockhash.BlockHash

BlockHash.CategorizeDef, BlockHash.GroupSpec, BlockHash.SortKey, BlockHash.TopNDef

Nested classes/interfaces inherited from interface org.elasticsearch.compute.aggregation.SeenGroupIds

SeenGroupIds.Empty, SeenGroupIds.Range

Field Summary

Fields

Modifier and Type	Field	Description
protected BlockHash	current
protected final int	emitBatchSize
protected final List<BlockHash.GroupSpec>	specs

Fields inherited from class org.elasticsearch.compute.aggregation.blockhash.BlockHash

blockFactory

Constructor Summary

Constructors

Constructor	Description
LongBytesRefAdaptiveBlockHash(List<BlockHash.GroupSpec> specs, BlockFactory blockFactory, int emitBatchSize, boolean reverseOutput)

Method Summary

Modifier and Type	Method	Description
final void	add(Page page, GroupingAggregatorFunction.AddInput addInput)	Add all values for the "group by" columns in the page to the hash and pass the ordinals to the provided GroupingAggregatorFunction.AddInput.
void	close()
final Block[]	getKeys(IntVector selected)	Returns an array of Blocks containing keys.
final ReleasableIterator<IntBlock>	lookup(Page page, ByteSizeValue targetBlockSize)	Lookup all values for the "group by" columns in the page to the hash and return an Iterator of the values.
boolean	migratedToPackedHash()	Whether this hash has migrated off the dictionary-aware vector-only fast path onto PackedValuesBlockHash.
final IntVector	nonEmpty()	The grouping ids that are not empty.
final int	numKeys()	The number of unique keys in the hash.
protected void	prepareAddInput(Page page)	Prepare the delegate block hash to add input from the given page.
protected void	prepareForLookup(Page page)	Prepare the delegate block hash for a lookup on the given page.
final BitArray	seenGroupIds(BigArrays bigArrays)	The grouping ids that have been seen already.
String	toString()

Methods inherited from class org.elasticsearch.compute.aggregation.blockhash.BlockHash

build, buildCategorizeBlockHash, buildPackedValuesBlockHash, hashOrdToGroup, hashOrdToGroupNullReserved

Methods inherited from class java.lang.Object

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

Field Details

specs

protected final List<BlockHash.GroupSpec> specs

emitBatchSize

protected final int emitBatchSize

current

protected BlockHash current

Constructor Details

LongBytesRefAdaptiveBlockHash

public LongBytesRefAdaptiveBlockHash(List<BlockHash.GroupSpec> specs,
 BlockFactory blockFactory,
 int emitBatchSize,
 boolean reverseOutput)

Method Details

prepareAddInput

protected void prepareAddInput(Page page)

Prepare the delegate block hash to add input from the given page.

prepareForLookup

protected void prepareForLookup(Page page)

Prepare the delegate block hash for a lookup on the given page.

migratedToPackedHash

public boolean migratedToPackedHash()

Whether this hash has migrated off the dictionary-aware vector-only fast path onto PackedValuesBlockHash. Once migrated, the hash never goes back. Intended for tests that need to pin "the fast path actually fired end-to-end" without scraping Object.toString().

add

public final void add(Page page,
 GroupingAggregatorFunction.AddInput addInput)

Description copied from class: BlockHash

Add all values for the "group by" columns in the page to the hash and pass the ordinals to the provided GroupingAggregatorFunction.AddInput.

This call will not Releasable.close() addInput.

Specified by:

add in class BlockHash

lookup

public final ReleasableIterator<IntBlock> lookup(Page page,
 ByteSizeValue targetBlockSize)

Description copied from class: BlockHash

Lookup all values for the "group by" columns in the page to the hash and return an Iterator of the values. The sum of Block.getPositionCount() for all blocks returned by the iterator will equal Page.getPositionCount() but will "target" a size of targetBlockSize.

The returned ReleasableIterator may retain a reference to Blocks inside the Page. Close it to release those references.

Specified by:

lookup in class BlockHash

getKeys

public final Block[] getKeys(IntVector selected)

Description copied from class: BlockHash

Returns an array of Blocks containing keys.

Specified by:

getKeys in class BlockHash

Parameters:

selected - The groupIds to include in the results. These are the same groupIds returned by BlockHash.nonEmpty() and fed into aggregations as part of BlockHash.add(org.elasticsearch.compute.data.Page, org.elasticsearch.compute.aggregation.GroupingAggregatorFunction.AddInput).

nonEmpty

public final IntVector nonEmpty()

Description copied from class: BlockHash

The grouping ids that are not empty. We use this because some block hashes reserve space for grouping ids and then don't end up using them. For example, BooleanBlockHash does this by always assigning false to 0 and true to 1. It's only after collection when we know if there actually were any true or false values received.

Specified by:

nonEmpty in class BlockHash

numKeys

public final int numKeys()

Description copied from class: BlockHash

The number of unique keys in the hash.

Specified by:

numKeys in class BlockHash

seenGroupIds

public final BitArray seenGroupIds(BigArrays bigArrays)

Description copied from interface: SeenGroupIds

The grouping ids that have been seen already. This BitArray is kept and mutated by the caller so make a copy if it's something you need your own copy of it.

Specified by:

seenGroupIds in interface SeenGroupIds

Specified by:

seenGroupIds in class BlockHash

close

public void close()

toString

public String toString()

Overrides:

toString in class Object