[jira] [Created] (TEZ-534) Add an InputFormat that combines original splits into groups

["Bikas Saha (JIRA)" <ji...@apache.org>]

Bikas Saha created TEZ-534:
------------------------------

             Summary: Add an InputFormat that combines original splits into groups
                 Key: TEZ-534
                 URL: https://issues.apache.org/jira/browse/TEZ-534
             Project: Apache Tez
          Issue Type: Bug
            Reporter: Bikas Saha
            Assignee: Bikas Saha


It would be useful if we could take the original input splits generated by a FileFormat and combine them to create a desired number of splits by combining them using location information. We get the advantage of using native logic to create the original splits as opposed to CombineFileInputFormat that does its own split calculation. However, native splits may be too large in number for efficient task execution. Combining them into a desired number of grouped splits will help.



--
This message was sent by Atlassian JIRA
(v6.1#6144)