[
https://issues.apache.org/jira/browse/PIG-2652?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13259324#comment-13259324
]
Dmitriy V. Ryaboy commented on PIG-2652:
----------------------------------------
Ok, not that simple. The adjuster messes with the inputs / outputs of the
limiting job in pretty complex ways, so one would have to unroll all of that
before running the pre-limit job.
Also, apparently one can't simply run the adjuster inside the
JobControlCompiler -- it does correctly add a new job, but that job fails with
{code}12/04/22 17:27:31 INFO mapred.TaskInProgress: Error from
attempt_20120422172532429_0004_m_000000_0:
org.apache.pig.backend.executionengine.ExecException: ERROR 2044: The type null
cannot be collected as a Key type
{code}
Moreover, the extra job is not accounted for in stats, progress, etc.
Looking at how we can do this better.
> Skew join and order by don't trigger reducer estimation
> -------------------------------------------------------
>
> Key: PIG-2652
> URL: https://issues.apache.org/jira/browse/PIG-2652
> Project: Pig
> Issue Type: Bug
> Reporter: Bill Graham
> Assignee: Bill Graham
> Fix For: 0.10.0, 0.9.3, 0.11
>
> Attachments: PIG-2652_1.patch, PIG-2652_2.patch, PIG-2652_3.patch,
> PIG-2652_3_10.patch, PIG-2652_4.patch, PIG-2652_5.patch
>
>
> If neither PARALLEL, default parallel or {{mapred.reduce.tasks}} are set, the
> number of reducers is not estimated based on input size for skew joins or
> order by. Instead, these jobs get only 1 reducer.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
