[jira] [Created] (SPARK-12400) Avoid writing a shuffle file if a partition has no output (empty)

["Reynold Xin (JIRA)" <ji...@apache.org>]

Reynold Xin created SPARK-12400:
-----------------------------------

             Summary: Avoid writing a shuffle file if a partition has no output (empty)
                 Key: SPARK-12400
                 URL: https://issues.apache.org/jira/browse/SPARK-12400
             Project: Spark
          Issue Type: Improvement
          Components: Shuffle
            Reporter: Reynold Xin


A Spark user was asking for automatic setting of # reducers. When I pushed for more, it turned out the problem for them is that 200 creates too many files, when most partitions are empty.

It seems like a simple thing we can do is to avoid creating shuffle files if a partition is empty.




--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org