Zhang Xiaoyang created HIVE-22567:
-------------------------------------

             Summary: Data loss when map join is off ,the result is diffrent 
when the number of reduce tasks is diffrent;
                 Key: HIVE-22567
                 URL: https://issues.apache.org/jira/browse/HIVE-22567
             Project: Hive
          Issue Type: Bug
    Affects Versions: 3.1.2, 3.1.1
         Environment: select count(1) from (
    select department_code 
    from tmp.tmp_mon
    where business_unit_code in (select business_unit_code from 
tmp.business_unit_config)
    group by department_code 
)tmp
            Reporter: Zhang Xiaoyang


I get diffrent results when the map join is off and the reduce tasks is 
diffrent !

the tmp.tmp_mon is a big table and tmp.business_unit_config has only 7 records;

when set hive.auto.convert.join=false,the result is diffrent when the number of 
the reduce tasks changed;

set mapred.reduce.tasks=1 the result seems right and when  set 
mapred.reduce.tasks=2 or other,the result is missing some data;

what can cause this ? 

 

 



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

Reply via email to