[HIVE-7736] improve the columns stats update speed for all the partitions of a table

• Type: Improvement
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 0.14.0
• Assignee: Pengcheng Xiong
• Reporter: Pengcheng Xiong
• Created: 14/Aug/14 22:32
• Updated: 13/Nov/14 19:44
• Resolved: 22/Aug/14 15:36

Description

The current implementation of columns stats update for all the partitions of a table takes a long time when there are thousands of partitions.

For example, on a given cluster, it took 600+ seconds to update all the partitions' columns stats for a table with 2 columns but 2000 partitions.

ANALYZE TABLE src_stat_part partition (partitionId) COMPUTE STATISTICS for columns;

We would like to improve the columns stats update speed for all the partitions of a table

Issue Links

• links to: RB request