Details
-
Bug
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
-
None
Description
When a target table is clustered into buckets, hive.enforce.bucketing will limit the number of reducers to match the number of buckets. It should take into account the number of partitions * number of buckets.
SET hive.exec.dynamic.partition.mode=nonstrict;
SET hive.enforce.bucketing = true;
SET hive.enforce.sorting = true;
INSERT INTO TABLE dest PARTITION(part) SELECT * FROM origin;
In this case, when dest has 5 buckets, the number of reducers will be 5 – if there are 10 partitions, the maximum reducers should be 50, not 5.