org.apache.druid.query.topn

Class TopNQueryEngine

java.lang.Object

org.apache.druid.query.topn.TopNQueryEngine

public class TopNQueryEngine
extends Object

Constructor Summary

Constructors

Constructor and Description
TopNQueryEngine(NonBlockingPool<ByteBuffer> bufferPool)

Method Summary

Modifier and Type	Method and Description
static boolean	canApplyExtractionInPost(TopNQuery query) ExtractionFn which are one to one may have their execution deferred until as late as possible, since the which value is used as the grouping key itself doesn't particularly matter.
Sequence<Result<TopNResultValue>>	query(TopNQuery query, StorageAdapter adapter, TopNQueryMetrics queryMetrics) Do the thing - process a StorageAdapter into a Sequence of TopNResultValue, with one of the fine TopNAlgorithm available chosen based on the type of column being aggregated.

Methods inherited from class java.lang.Object

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

Constructor Detail

TopNQueryEngine

public TopNQueryEngine(NonBlockingPool<ByteBuffer> bufferPool)

Method Detail

query

public Sequence<Result<TopNResultValue>> query(TopNQuery query,
                                               StorageAdapter adapter,
                                               @Nullable
                                               TopNQueryMetrics queryMetrics)

Do the thing - process a StorageAdapter into a Sequence of TopNResultValue, with one of the fine TopNAlgorithm available chosen based on the type of column being aggregated. The algorithm provides a mapping function to process rows from the adapter Cursor to apply AggregatorFactory and create or update TopNResultValue

canApplyExtractionInPost

public static boolean canApplyExtractionInPost(TopNQuery query)

ExtractionFn which are one to one may have their execution deferred until as late as possible, since the which value is used as the grouping key itself doesn't particularly matter. For top-n, this method allows the query to be transformed in TopNQueryQueryToolChest.preMergeQueryDecoration(org.apache.druid.query.QueryRunner<org.apache.druid.query.Result<org.apache.druid.query.topn.TopNResultValue>>) to strip off the ExtractionFn on the broker, so that a more optimized algorithm (e.g. PooledTopNAlgorithm) can be chosen for processing segments, and then added back and evaluated against the final merged result sets on the broker via TopNQueryQueryToolChest.postMergeQueryDecoration(org.apache.druid.query.QueryRunner<org.apache.druid.query.Result<org.apache.druid.query.topn.TopNResultValue>>).