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

Posted to dev@hive.apache.org by "Gunther Hagleitner (JIRA)" <ji...@apache.org> on 2013/10/05 01:16:42 UTC

[jira] [Commented] (HIVE-5451) Use IOContext instead of HADOOPMAPFILENAME in MapOperator setup for Tez

    [ https://issues.apache.org/jira/browse/HIVE-5451?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13786740#comment-13786740 ] 

Gunther Hagleitner commented on HIVE-5451:
------------------------------------------

Sorry - I meant branch. I committed to Tez branch *not* trunk.

> Use IOContext instead of HADOOPMAPFILENAME in MapOperator setup for Tez
> -----------------------------------------------------------------------
>
>                 Key: HIVE-5451
>                 URL: https://issues.apache.org/jira/browse/HIVE-5451
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Gunther Hagleitner
>            Assignee: Gunther Hagleitner
>             Fix For: tez-branch
>
>         Attachments: HIVE-5451.1.patch
>
>
> In mapop we use a conf variable to determine input file, everywhere else we rely on IOContext. In tez the problem is that the recordreader and processor have different confs. So they can't communicate via job conf.



--
This message was sent by Atlassian JIRA
(v6.1#6144)