Details
-
Sub-task
-
Status: Open
-
Major
-
Resolution: Unresolved
-
None
-
None
-
None
Description
In the plan of a query, if there exists MapReduce jobs created for aggregation operators or join operators which share one or more input tables and use the same partitioning keys to shuffle the data, these jobs can be merged into a single MapReduce job. Shared tables will be only scanned once.
Attachments
Issue Links
- is blocked by
-
HIVE-2206 add a new optimizer for query correlation discovery and optimization
- Closed