[HIVE-13260] ReduceSinkDeDuplication throws exception when pRS key is empty

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.0.0
• Fix Version/s: 2.1.0
• Component/s: Logical Optimizer
• Assignee: Pengcheng Xiong
• Reporter: Pengcheng Xiong
• Created: 10/Mar/16 21:12
• Updated: 13/Aug/20 08:42
• Resolved: 16/Mar/16 17:08

Description

Steps to reproduce:

set hive.mapred.mode=nonstrict;
set hive.cbo.enable=false;

set hive.map.aggr=false;

set hive.groupby.skewindata=false;
set mapred.reduce.tasks=31;

select compute_stats(a,16),compute_stats(b,16),compute_stats(c,16),compute_stats(d,16)
from
(
select
  avg(DISTINCT substr(src.value,5)) as a,
  max(substr(src.value,5)) as b,
  variance(substr(src.value,5)) as c,
  var_samp(substr(src.value,5)) as d
 from src)subq;

Issue Links

• blocks: HIVE-11160 Auto-gather column stats
• causes: HIVE-18284 NPE when inserting data with 'distribute by' clause with dynpart sort optimization
• links to: RB