Package org.apache.druid.query.groupby.epinephelinae

Class GroupByMergingQueryRunner

java.lang.Object

org.apache.druid.query.groupby.epinephelinae.GroupByMergingQueryRunner

All Implemented Interfaces:

QueryRunner<ResultRow>

public class GroupByMergingQueryRunner
extends Object
implements QueryRunner<ResultRow>

Class that knows how to merge a collection of groupBy QueryRunner objects, called queryables, using a buffer provided by mergeBufferPool and a parallel executor provided by exec. Outputs a fully aggregated stream of ResultRow objects. Does not apply post-aggregators. The input queryables are expected to come from a GroupByQueryEngine. This code primarily runs on data servers like Historicals and Realtime Tasks. This can also run on Brokers, if the query is operating on local data sources, like inlined data, where the broker itself acts like a data server This class has some resemblance to GroupByRowProcessor. See the javadoc of that class for a discussion of similarities and differences. Used by GroupingEngine.mergeRunners(QueryProcessingPool, Iterable)

Constructor Summary

Constructors

Constructor

Description

GroupByMergingQueryRunner(GroupByQueryConfig config, DruidProcessingConfig processingConfig, QueryProcessingPool queryProcessingPool, QueryWatcher queryWatcher, Iterable<QueryRunner<ResultRow>> queryables, GroupByResourcesReservationPool groupByResourcesReservationPool, int concurrencyHint, int mergeBufferSize, com.fasterxml.jackson.databind.ObjectMapper spillMapper, String processingTmpDir, GroupByStatsProvider groupByStatsProvider)

Method Summary

Modifier and Type	Method	Description
Sequence<ResultRow>	run(QueryPlus<ResultRow> queryPlus, ResponseContext responseContext)	Runs the given query and returns results in a time-ordered sequence.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.druid.query.QueryRunner

run

Constructor Details

GroupByMergingQueryRunner

public GroupByMergingQueryRunner(GroupByQueryConfig config,
 DruidProcessingConfig processingConfig,
 QueryProcessingPool queryProcessingPool,
 QueryWatcher queryWatcher,
 Iterable<QueryRunner<ResultRow>> queryables,
 GroupByResourcesReservationPool groupByResourcesReservationPool,
 int concurrencyHint,
 int mergeBufferSize,
 com.fasterxml.jackson.databind.ObjectMapper spillMapper,
 String processingTmpDir,
 GroupByStatsProvider groupByStatsProvider)

Method Details

run

public Sequence<ResultRow> run(QueryPlus<ResultRow> queryPlus,
 ResponseContext responseContext)

Description copied from interface: QueryRunner

Runs the given query and returns results in a time-ordered sequence.

Specified by:

run in interface QueryRunner<ResultRow>