[HIVE-19481] Tablesample uses incorrect logic to pick files corresponding to buckets.

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.0.0
• Fix Version/s: 3.1.0, 4.0.0-alpha-1
• Component/s: Hive
• Assignee: Deepak Jaiswal
• Reporter: Steve Yeom
• Created: 09/May/18 21:44
• Updated: 21/Oct/22 06:59
• Resolved: 27/Jun/18 06:32

Description

Ran "mvn test -Dtest=TestMiniLlapLocalCliDriver -Dqfile=sample10.q " after changing the table to be
insert-only transactional.

The following queries returns couple of rows whereis no row results returns for non-ACID table.
query: select ds, count(1) from srcpartbucket tablesample (bucket 2 out of 4 on key) where ds is not null group by ds ORDER BY ds ASC

2008-04-08 14
2008-04-09 14
..
query: select ds, count(1) from srcpartbucket tablesample (bucket 1 out of 2 on key) where ds is not null group by ds ORDER BY ds ASC

2008-04-08 4
2008-04-09 4

Issue Links

• links to: RB