Details
-
Improvement
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
Description
Currently, the following parameter: BYTESPERREDUCER("hive.exec.reducers.bytes.per.reducer", (long) (1000 * 1000 * 1000)),
decides the number of reducers for a given query - it is same for all map-reduce jobs of the query. It would be good to give more flexibility to the
user, and provide a way so that they can determine number of reducers for each map-reduce job
Attachments
Issue Links
- relates to
-
HIVE-17004 Calculating Number Of Reducers Looks At All Files
- Open