Description
SMB joins are used wherever the tables are sorted and bucketed. It's a map-side join. The join boils down to just merging the already sorted tables, allowing this operation to be faster than an ordinary map-join.
The task is to research and support the conversion from regular SMB join to SMB map join for Spark execution engine.
Attachments
Attachments
Issue Links
- incorporates
-
HIVE-9192 One-pass SMB Optimizations [Spark Branch]
- Resolved
- is depended upon by
-
HIVE-8216 auto_smb_mapjoin_14.q failed test with exception. [Spark Branch]
- Resolved
- is related to
-
HIVE-8639 Convert SMBJoin to MapJoin [Spark Branch]
- Resolved
-
HIVE-8640 Support hints of SMBJoin [Spark Branch]
- Resolved
- links to