spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Apache Spark (Jira)" <j...@apache.org>
Subject [jira] [Commented] (SPARK-32761) Planner error when aggregating multiple distinct Constant columns
Date Tue, 01 Sep 2020 05:29:00 GMT

    [ https://issues.apache.org/jira/browse/SPARK-32761?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17188146#comment-17188146
] 

Apache Spark commented on SPARK-32761:
--------------------------------------

User 'linhongliu-db' has created a pull request for this issue:
https://github.com/apache/spark/pull/29607

> Planner error when aggregating multiple distinct Constant columns
> -----------------------------------------------------------------
>
>                 Key: SPARK-32761
>                 URL: https://issues.apache.org/jira/browse/SPARK-32761
>             Project: Spark
>          Issue Type: Bug
>          Components: SQL
>    Affects Versions: 3.0.0
>            Reporter: Linhong Liu
>            Priority: Major
>
> SELECT COUNT(DISTINCT 2), COUNT(DISTINCT 2, 3) will trigger this bug.
> The problematic code is:
>  
> {code:java}
> val distinctAggGroups = aggExpressions.filter(_.isDistinct).groupBy { e =>
>   val unfoldableChildren = e.aggregateFunction.children.filter(!_.foldable).toSet
>   if (unfoldableChildren.nonEmpty) {
>     // Only expand the unfoldable children
>      unfoldableChildren
>   } else {
>     // If aggregateFunction's children are all foldable
>     // we must expand at least one of the children (here we take the first child),
>     // or If we don't, we will get the wrong result, for example:
>     // count(distinct 1) will be explained to count(1) after the rewrite function.
>     // Generally, the distinct aggregateFunction should not run
>     // foldable TypeCheck for the first child.
>     e.aggregateFunction.children.take(1).toSet
>   }
> }
> {code}



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org


Mime
View raw message