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

Class GroupByRowProcessor

java.lang.Object

org.apache.druid.query.groupby.epinephelinae.GroupByRowProcessor

public class GroupByRowProcessor
extends Object

Utility class that knows how to do higher-level groupBys: i.e. group a Sequence of ResultRow originating from a subquery. It uses a buffer provided by a GroupByQueryResources. The output rows may not be perfectly grouped and will not have PostAggregators applied, so they should be fed into GroupingEngine.mergeResults(org.apache.druid.query.QueryRunner<org.apache.druid.query.groupby.ResultRow>, org.apache.druid.query.groupby.GroupByQuery, org.apache.druid.query.context.ResponseContext). This class has two primary uses: processing nested groupBys, and processing subtotals. This class has some similarity to GroupByMergingQueryRunnerV2, but is different enough that it deserved to be its own class. Some common code between the two classes is in RowBasedGrouperHelper.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static interface	GroupByRowProcessor.ResultSupplier

Method Summary

Modifier and Type	Method	Description
static GroupByRowProcessor.ResultSupplier	process(GroupByQuery query, GroupByQuery subquery, Sequence<ResultRow> rows, GroupByQueryConfig config, DruidProcessingConfig processingConfig, GroupByQueryResources resource, com.fasterxml.jackson.databind.ObjectMapper spillMapper, String processingTmpDir, int mergeBufferSize)	Process the input of sequence "rows" (output by "subquery") based on "query" and returns a GroupByRowProcessor.ResultSupplier.

Methods inherited from class java.lang.Object

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

Method Detail

process

public static GroupByRowProcessor.ResultSupplier process(GroupByQuery query,
                                                         GroupByQuery subquery,
                                                         Sequence<ResultRow> rows,
                                                         GroupByQueryConfig config,
                                                         DruidProcessingConfig processingConfig,
                                                         GroupByQueryResources resource,
                                                         com.fasterxml.jackson.databind.ObjectMapper spillMapper,
                                                         String processingTmpDir,
                                                         int mergeBufferSize)

Process the input of sequence "rows" (output by "subquery") based on "query" and returns a GroupByRowProcessor.ResultSupplier. In addition to grouping using dimensions and metrics, it will also apply filters (both DimFilter and interval filters). The input sequence is processed synchronously with the call to this method, and result iteration happens lazy upon calls to the GroupByRowProcessor.ResultSupplier. Make sure to close it when you're done.