[HIVE-2119] Optimizer on partition field

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.8.0
• Assignee: Unassigned
• Reporter: Ron Bodkin
• Created: 18/Apr/11 23:42
• Updated: 02/May/13 02:29
• Resolved: 31/Oct/11 17:38

Description

The optimizer doesn't use partition metadata, e.g.,

select max(dt) from big_partitioned_table;

Instead of just returning the biggest date in the table, it runs a full table scan. This kind of expression can also be useful in more interesting cases like:

select count from big_partitioned_table a join (select max(dt) m from big_partitioned_table) b on a.dt=b.m;

Issue Links

• depends upon: HIVE-1003 optimize metadata only queries