[HIVE-25580] Increase the performance of getTableColumnStatistics and getPartitionColumnStatistics

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 4.0.0-alpha-1
• Component/s: Standalone Metastore
• Labels: pull-request-available
• Assignee: Peter Vary
• Reporter: Peter Vary
• Created: 30/Sep/21 12:48
• Updated: 17/Nov/22 08:48
• Resolved: 11/Oct/21 11:23

Description

When the PART_COL_STATS table contains high number of rows the getTableColumnStatistics and getPartitionColumnStatistics response time increases.

The root cause is the full table scan for the jdbc query below:

2021-09-27 13:22:44,218 DEBUG DataNucleus.Datastore.Native: [pool-6-thread-199]: SELECT DISTINCT "A0"."ENGINE" FROM "PART_COL_STATS" "A0"
2021-09-27 13:22:50,569 DEBUG DataNucleus.Datastore.Retrieve: [pool-6-thread-199]: Execution Time = 6351 ms 

The time spent in here:

      query = pm.newQuery(MPartitionColumnStatistics.class);
      query.setResult("DISTINCT engine");
      Collection names = (Collection) query.execute();

We might get a better performance if we limit the query range based on the cat/db/table.

Issue Links

• links to: GitHub Pull Request #2692