[IMPALA-20] Aggregate of a subquery result set returns wrong results if the subquery contains a 'limit' and data is distributed across multiple nodes

• Type: Bug
• Status: Resolved
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: Impala 0.3
• Fix Version/s: Impala 0.7
• Assignee: Marcel Kinard
• Reporter: Lenni Kuff
• Created: 17/Jan/13 00:08
• Updated: 20/Dec/15 00:04
• Resolved: 10/Mar/13 01:29

Description

Aggregate of a subquery result set returns wrong results if the subquery contains a 'limit' clause and data is distributed across multiple nodes. From the query plan, it looks like we are just summing the results from each slave.

Example, if the data spread across 3 nodes (expected result is 10):

> select count(*) from (select * from tpch.lineitem limit 10) p
Query finished, fetching results ...
30
Returned 1 row(s) in 0.08s

Plan

 
  UNPARTITIONED
  AGGREGATE
  OUTPUT: SUM(<slot 32>)
  GROUP BY:
  TUPLE IDS: 2
    EXCHANGE (2)
      TUPLE IDS: 2

Plan Fragment 1
  RANDOM
  STREAM DATA SINK
    EXCHANGE ID: 2
    UNPARTITIONED

  AGGREGATE
  OUTPUT: COUNT(*)
  GROUP BY:
  TUPLE IDS: 2
    SCAN HDFS table=tpch.lineitem #partitions=1 size=718.94MB (0)
      LIMIT: 10
      TUPLE IDS: 0