[jira] Resolved: (HADOOP-587) SequenceFile sort should use quicksort instead of merge sort for sorting runs.

["Owen O'Malley (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HADOOP-587?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Owen O'Malley resolved HADOOP-587.
----------------------------------

    Resolution: Duplicate

Fixed by HADOOP-2919.

> SequenceFile sort should use quicksort instead of merge sort for sorting runs.
> ------------------------------------------------------------------------------
>
>                 Key: HADOOP-587
>                 URL: https://issues.apache.org/jira/browse/HADOOP-587
>             Project: Hadoop Core
>          Issue Type: Improvement
>          Components: mapred
>            Reporter: Runping Qi
>            Assignee: Owen O'Malley
>            Priority: Minor
>
> I noticed that Hadoop uses mergesort for sorting the runs in the sequence file SortPass class. 
> I think quicksort should be faster and should be used instead. 
> I also noticed that Java's Arrays.sort() static functions for non-primitive element types also use merge sort. 
> Are there any subtle reasons why not to use quicksort?
> If no, we should implement out quicksort. Otherwise, we should use Java's sort, unless we believe our mergesort is better.

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