Details
-
Improvement
-
Status: Closed
-
Major
-
Resolution: Incomplete
-
None
-
None
-
None
Description
Our mapred implementation is pretty meager compared to the code in mapreduce. Notably, it does not support a user-provided Scan instance for mapreduce jobs run over hbase tables. The only thing we offer is column family filtering. For applications to support predicate pushdown and enjoy efficient HBase access, either they transition their entire codebase over to mapreduce, or we offer this functionality.
Attachments
Issue Links
- is depended upon by
-
HIVE-7058 Cleanup HiveHBase*InputFormat
- Open
- is related to
-
HBASE-11137 Add mapred.TableSnapshotInputFormat
- Closed
-
HIVE-6584 Add HiveHBaseTableSnapshotInputFormat
- Closed
- relates to
-
HBASE-11179 API parity between mapred and mapreduce
- Closed