Details
-
New Feature
-
Status: Closed
-
Major
-
Resolution: Fixed
-
None
-
None
-
None
-
Reviewed
Description
Currently the Bloom Filter from BuildBloom has to be stored to HDFS to be able to be used in Bloom UDF. Most of the time the bloom filter is not reused and so have to be deleted after the end of the script. The load/store also forces multiple DAGs. If it was passed as a scalar, then it would be simpler and more efficient.
Attachments
Attachments
Issue Links
- duplicates
-
PIG-2348 Bloom should be able to take a relation or a file
- Resolved