hadoop-hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Carl Steinbach (JIRA)" <j...@apache.org>
Subject [jira] Updated: (HIVE-1404) Allow UDAFs to turn off map-side aggregation
Date Mon, 14 Jun 2010 23:02:13 GMT

     [ https://issues.apache.org/jira/browse/HIVE-1404?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Carl Steinbach updated HIVE-1404:
---------------------------------

    Affects Version/s: 0.5.0
          Component/s: Query Processor
                       UDF

> Allow UDAFs to turn off map-side aggregation
> --------------------------------------------
>
>                 Key: HIVE-1404
>                 URL: https://issues.apache.org/jira/browse/HIVE-1404
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Query Processor, UDF
>    Affects Versions: 0.5.0
>            Reporter: Jonathan Chang
>            Priority: Minor
>
> Some UDAFs such as COLLECT will not work with map-side aggregation (because each mapper
must hold all of its contents in memory).   To use such a function at the moment, one has
to remember to set hive.map.aggr.  This has two problems:
> 1.) The user has to remember to set this flag or else bad things will happen.
> 2.) The flag applies to the entire query plan rather than just the block containing the
COLLECT statement.
> It would be good to have a bit in the UDAF which would tell the compiler to turn off
map-side aggregation for just the block containing the UDAF.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message