You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@hive.apache.org by "Prasanth J (JIRA)" <ji...@apache.org> on 2014/04/23 00:22:16 UTC

[jira] [Updated] (HIVE-6956) Duplicate partitioning column for union when dynamic partition sort optimization is enabled

     [ https://issues.apache.org/jira/browse/HIVE-6956?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Prasanth J updated HIVE-6956:
-----------------------------

    Attachment: HIVE-6956.1.patch

The test diffs are valid since SEL->RS->EX->SEL->FS is same as SEL->RS->EX->FS.. The diffs have the redundant SEL in reducer removed after changing the graph walker.

> Duplicate partitioning column for union when dynamic partition sort optimization is enabled
> -------------------------------------------------------------------------------------------
>
>                 Key: HIVE-6956
>                 URL: https://issues.apache.org/jira/browse/HIVE-6956
>             Project: Hive
>          Issue Type: Bug
>    Affects Versions: 0.13.0, 0.14.0
>            Reporter: Prasanth J
>            Assignee: Prasanth J
>              Labels: optimization
>         Attachments: HIVE-6956.1.patch
>
>
> HIVE-6455 optimization uses PreOrder graph walker resulting in double invocation of SortedDynPartitionOptimizer(). Because of this there are duplicate partitioning columns in RSOp for few test cases (union_remove_17.q and similar test cases..)..
> HIVE-6455 optimization does not actually need a PreOrder graph walker. The fix is to revert it to DefaultGraphWalker.



--
This message was sent by Atlassian JIRA
(v6.2#6252)