You are viewing a plain text version of this content. The canonical link for it is here.

Posted to dev@pig.apache.org by "Laukik Chitnis (JIRA)" <ji...@apache.org> on 2009/02/06 00:35:59 UTC

[jira] Created: (PIG-657) splitsize is ignored in PigInputFormat

splitsize is ignored in PigInputFormat
--------------------------------------

                 Key: PIG-657
                 URL: https://issues.apache.org/jira/browse/PIG-657
             Project: Pig
          Issue Type: Bug
            Reporter: Laukik Chitnis


The way to control the number of mappers in Hadoop has been to specify a mapred.min.split.size parameter in the job conf. For eg.  mapred.min.split.size=1073741824,mapred.map.tasks=10

However, even if this parameter is specified, Pig creates the number of mappers depending only on the number of blocks in the file. This is because the parameter is not used in the PigInputFormat.

The parameter can actually be extracted from the job conf object. So, one way of doing this would be to pass an handle to the job conf object to the PigInputFormat or the custom slicer.




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

[jira] Resolved: (PIG-657) splitsize is ignored in PigInputFormat

Posted by "Olga Natkovich (JIRA)" <ji...@apache.org>.

     [ https://issues.apache.org/jira/browse/PIG-657?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Olga Natkovich resolved PIG-657.
--------------------------------

    Resolution: Fixed

This is resolved with Pig 0.7

> splitsize is ignored in PigInputFormat
> --------------------------------------
>
>                 Key: PIG-657
>                 URL: https://issues.apache.org/jira/browse/PIG-657
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Laukik Chitnis
>
> The way to control the number of mappers in Hadoop has been to specify a mapred.min.split.size parameter in the job conf. For eg.  mapred.min.split.size=1073741824,mapred.map.tasks=10
> However, even if this parameter is specified, Pig creates the number of mappers depending only on the number of blocks in the file. This is because the parameter is not used in the PigInputFormat.
> The parameter can actually be extracted from the job conf object. So, one way of doing this would be to pass an handle to the job conf object to the PigInputFormat or the custom slicer.

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