[ https://issues.apache.org/jira/browse/SPARK-27290?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16803737#comment-16803737 ]
Xiaoju Wu commented on SPARK-27290: ----------------------------------- [~ekoifman] HashAggregate can not benefit from sorted input but SortAggregate can. But SortAggregate will require its input sorted by columns of group-by, and will enforce the SortExec in the logic of EnsureRequirements. So the user-written sort is useless. > remove unneed sort under Aggregate > ---------------------------------- > > Key: SPARK-27290 > URL: https://issues.apache.org/jira/browse/SPARK-27290 > Project: Spark > Issue Type: Improvement > Components: SQL > Affects Versions: 2.4.0 > Reporter: Xiaoju Wu > Priority: Minor > > I saw some tickets to remove unneeded sort in plan while I think there's > another case in which sort is redundant: > Sort just under an non-orderPreserving node is redundant, for example: > select count(*) from (select a1 from A order by a2); > +- Aggregate > +- Sort > +- FileScan parquet > But one of the existing test cases is conflict with this example: > test("sort should not be removed when there is a node which doesn't guarantee > any order") > { val orderedPlan = testRelation.select('a, 'b).orderBy('a.asc) val > groupedAndResorted = orderedPlan.groupBy('a)(sum('a)).orderBy('a.asc) val > optimized = Optimize.execute(groupedAndResorted.analyze) val correctAnswer > = groupedAndResorted.analyze comparePlans(optimized, correctAnswer) } > Why is it designed like this? In my opinion, since Aggregate won't pass up > the ordering, the below Sort is useless. -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org For additional commands, e-mail: issues-h...@spark.apache.org