hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Steve Loughran (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-4827) Increase hash quality of HashPartitioner
Date Thu, 29 Nov 2012 14:47:00 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-4827?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13506505#comment-13506505

Steve Loughran commented on MAPREDUCE-4827:

This looks good to me
 # as doug says, this could be a regression, perhaps a {{BetterHashPartitioner}} is needed,
or make this an option.
 # we cannot put code from the JDK into the tree -or code that looks exactly like it, as we
don't want another Oracle related lawsuit. That means we need to find some tangible reference
-such as the page in Knuth, and we have to work it out from there by someone who hasn't looked
at the HashMap code.
This is being over cautious, but given Oracle sued google over a max function we have to show
that diligence.
> Increase hash quality of HashPartitioner
> ----------------------------------------
>                 Key: MAPREDUCE-4827
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-4827
>             Project: Hadoop Map/Reduce
>          Issue Type: Improvement
>            Reporter: Radim Kolar
>         Attachments: betterhash1.txt
> hash partitioner is using object.hashCode() for splitting keys into partitions. This
results in bad distributions because hashCode() quality is poor. 
> These hashCode() functions are sometimes written by hand (very poor quality) and sometimes
generated from by commons lang code (poor quality). Applying some transformation on top of
hashCode() provides better distribution.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

View raw message