[jira] [Commented] (SPARK-11028) When planning queries without partial aggregation support, we should try to use TungstenAggregate.

["Josh Rosen (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-11028?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14951210#comment-14951210 ] 

Josh Rosen commented on SPARK-11028:
------------------------------------

[~yhuai], if we fix SPARK-10992 first then will we still need to do this? Will it still be the case that _some_ HiveUDAFs don't support partial aggregation, requiring this?

> When planning queries without partial aggregation support, we should try to use TungstenAggregate.
> --------------------------------------------------------------------------------------------------
>
>                 Key: SPARK-11028
>                 URL: https://issues.apache.org/jira/browse/SPARK-11028
>             Project: Spark
>          Issue Type: Sub-task
>          Components: SQL
>            Reporter: Yin Huai
>
> With SPARK-11017, we can run DeclarativeAggregate Functions in TungstenAggregate. So, when we plan queries having functions that do not support partial aggregation, we can use TungstenAggregate whenever possible. The reason that we only use SortBasedAggregate is that HiveUDAF is the only function that does not support partial aggregation and it is a DeclarativeAggregate function.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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