[jira] [Updated] (CRUNCH-51) PCollection#sort relies on using a single reducer for total order sorting

["Rahul Sharma (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/CRUNCH-51?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Rahul Sharma updated CRUNCH-51:
-------------------------------

    Attachment: 0001-CRUNCH-51-Total-Order-Sort.patch

Reservoir sampling has been used to select a sample and then created a partitioning file for all the required keys. Also TotalOrderPartitioner has been copied from hadoop and modified to make it work for avro data also. I have taken a shot at fixing this one for avro and writable data both.
This can only be tested on a clustered environment, so I could not make out a test case for the same. 
Gabriel, can you have a look at this one ?
                
> PCollection#sort relies on using a single reducer for total order sorting
> -------------------------------------------------------------------------
>
>                 Key: CRUNCH-51
>                 URL: https://issues.apache.org/jira/browse/CRUNCH-51
>             Project: Crunch
>          Issue Type: Improvement
>    Affects Versions: 0.3.0
>            Reporter: Gabriel Reid
>         Attachments: 0001-CRUNCH-51-Total-Order-Sort.patch
>
>
> The total-order sorting provided by the Sort class (and therefore PCollection#sort) relies on using a single reducer in order to provide total-order sorting. This is very inefficient for large datasets, and should be replaced with a total order partitioner instead.
> For more information, see CRUNCH-23 (and possibly also MAPREDUCE-4574).

--
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