[
https://issues.apache.org/jira/browse/SPARK-14830?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15253047#comment-15253047
]
Apache Spark commented on SPARK-14830:
--------------------------------------
User 'dongjoon-hyun' has created a pull request for this issue:
https://github.com/apache/spark/pull/12590
> Add RemoveRepetitionFromGroupExpressions optimizer
> --------------------------------------------------
>
> Key: SPARK-14830
> URL: https://issues.apache.org/jira/browse/SPARK-14830
> Project: Spark
> Issue Type: Improvement
> Components: Optimizer
> Reporter: Dongjoon Hyun
>
> This issue aims to optimize GroupExpressions by removing repeating
> expressions.
> **Before**
> {code}
> scala> sql("select a from (select explode(array(1,2)) a) T group by a, a,
> a").explain()
> == Physical Plan ==
> WholeStageCodegen
> : +- TungstenAggregate(key=[a#5,a#5,a#5], functions=[], output=[a#5])
> : +- INPUT
> +- Exchange hashpartitioning(a#5, a#5, a#5, 200), None
> +- WholeStageCodegen
> : +- TungstenAggregate(key=[a#5,a#5,a#5], functions=[],
> output=[a#5,a#5,a#5])
> : +- INPUT
> +- Generate explode([1,2]), false, false, [a#5]
> +- Scan OneRowRelation[]
> {code}
> **After**
> {code}
> scala> sql("select a from (select explode(array(1,2)) a) T group by a, a,
> a").explain()
> == Physical Plan ==
> WholeStageCodegen
> : +- TungstenAggregate(key=[a#5], functions=[], output=[a#5])
> : +- INPUT
> +- Exchange hashpartitioning(a#5, 200), None
> +- WholeStageCodegen
> : +- TungstenAggregate(key=[a#5], functions=[], output=[a#5])
> : +- INPUT
> +- Generate explode([1,2]), false, false, [a#5]
> +- Scan OneRowRelation[]
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
