hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Rui Li (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-16659) Query plan should reflect hive.spark.use.groupby.shuffle
Date Fri, 30 Jun 2017 15:40:01 GMT

    [ https://issues.apache.org/jira/browse/HIVE-16659?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16070297#comment-16070297
] 

Rui Li commented on HIVE-16659:
-------------------------------

Thanks [~libing] for working on this. Patch looks good overall. Can you create an RB for it?
And a side note is, we usually base the patch against master. If you want to develop for a
specific branch, you should name your patch accordingly (refer to this [wiki|https://cwiki.apache.org/confluence/display/Hive/HowToContribute#HowToContribute-CreatingaPatch]).

I think it's OK to remove the RepartitionShuffler, as it's effectively the same as SortByShuffler
w/o total order. [~csun], [~xuefuz] please let me know if you think otherwise.

> Query plan should reflect hive.spark.use.groupby.shuffle
> --------------------------------------------------------
>
>                 Key: HIVE-16659
>                 URL: https://issues.apache.org/jira/browse/HIVE-16659
>             Project: Hive
>          Issue Type: Bug
>          Components: Spark
>            Reporter: Rui Li
>            Assignee: Bing Li
>         Attachments: HIVE-16659.1.patch
>
>
> It's useful to show the shuffle type used in the query plan. Currently it shows "GROUP"
no matter what we set for hive.spark.use.groupby.shuffle.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message