You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Ashutosh Chauhan (JIRA)" <ji...@apache.org> on 2016/03/31 18:16:25 UTC
[jira] [Commented] (HIVE-13382) Dynamic partitioning for INSERT
OVERWRITE Query is slow
[ https://issues.apache.org/jira/browse/HIVE-13382?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15220112#comment-15220112 ]
Ashutosh Chauhan commented on HIVE-13382:
-----------------------------------------
I have done patch series for this : HIVE-12908 HIVE-12988 HIVE-12907 & HIVE-12897 Unfortunately all of them are not in any release yet. You may want to backport to your deployment. There should be a substantial speedup for this.
> Dynamic partitioning for INSERT OVERWRITE Query is slow
> -------------------------------------------------------
>
> Key: HIVE-13382
> URL: https://issues.apache.org/jira/browse/HIVE-13382
> Project: Hive
> Issue Type: Bug
> Components: Hive
> Reporter: Lin Liu
> Priority: Blocker
>
> Recently we upgraded Hive from 0.13.1 to Hive 1.2.1.
> Then we found that one of our production queries that updates the data for a multi-level partitioned table using INSERT OVERWRITE becomes dramatically slower than in Hive0.13.1 version (e.g. from 2 hours to more than 14 hours and slower and slower).
> Note that the MR jobs finished pretty quickly, but the move stage took a very long time.
> Currently we are using Hive 1.2.1 and HIVE-11940 has been applied.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)