[HIVE-9045] Choosing right preference between map join and bucket map join [Spark Branch]

• Type: Sub-task
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: Spark
• Assignee: Unassigned
• Reporter: Xuefu Zhang
• Created: 09/Dec/14 00:10
• Updated: 09/Dec/14 00:11

Description

Currently We convert a join to bucket map join only after it's converted to map join. Supposedly, however, bucket map join has a smaller memory print, so a join which doesn't fit to map join might be suitable for a bucket map join. It would be good if we can take this into consideration when applying join optimizations.