Milind Bhandarkar commented on PIG-840:

Ideally, a slicer should only be a thin wrapper on top of InputSplit, so that 
all the different kinds of splits already part of hadoop (such as one-line 
input split, multi-file input split, default file-input split) etc are all 
supported out of the box.

> default slicer should support Hadoop's standard way to setup split size
> -----------------------------------------------------------------------
>                 Key: PIG-840
>                 URL: https://issues.apache.org/jira/browse/PIG-840
>             Project: Pig
>          Issue Type: Bug
>            Reporter: Olga Natkovich
> It should honor mapre.min.split.size rather than using its own var of 
> pig.overrideBlockSize

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

Reply via email to