[
https://issues.apache.org/jira/browse/HIVE-25838?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
dinesh updated HIVE-25838:
--------------------------
Description:
Hive SQL's on empty partitions giving no result instead of 0 rows or actual
value. For example -
--Create external Table
1) Create external table test_tbl ( name string) partitioned by ( company
string, processdate string) stored as orc location '/my/some/random/location';
– Add partion
2) Alter table test_tbl add partition ( company='aquaifer',
processdate='20220101');
– Execute following SQL's which returns no records.
3) select count( * ) , processdate from test_tbl group by processdate ;
4) select max(processdate ) from test_tbl ;
Same SQL (#3 & #4 above) , when execute with SPARK, returns '0' count and
'20220101' respectively.
was:
Hive SQL's on empty partitions giving no result instead of 0 rows or actual
value. For example -
--Create external Table
1) Create external table test_tbl ( name string) partitioned by ( company
string, processdate string) stored as orc location '/my/some/random/location';
-- Add partion
2) Alter table test_tbl add partition ( company='aquaifer',
processdate='20220101');
-- Execute following SQL's which returns no records.
3) select count(*), processdate from test_tbl group by processdate ;
4) select max(processdate ) from test_tbl ;
Same SQL (#3 & #4 above) , when execute with SPARK, returns '0' count and
'20220101' respectively.
> Hive SQL using TEZ as execution engine not giving result on empty partition
> ---------------------------------------------------------------------------
>
> Key: HIVE-25838
> URL: https://issues.apache.org/jira/browse/HIVE-25838
> Project: Hive
> Issue Type: Bug
> Affects Versions: 3.1.0
> Reporter: dinesh
> Priority: Major
>
> Hive SQL's on empty partitions giving no result instead of 0 rows or actual
> value. For example -
> --Create external Table
> 1) Create external table test_tbl ( name string) partitioned by ( company
> string, processdate string) stored as orc location '/my/some/random/location';
> – Add partion
> 2) Alter table test_tbl add partition ( company='aquaifer',
> processdate='20220101');
>
> – Execute following SQL's which returns no records.
> 3) select count( * ) , processdate from test_tbl group by processdate ;
> 4) select max(processdate ) from test_tbl ;
>
> Same SQL (#3 & #4 above) , when execute with SPARK, returns '0' count and
> '20220101' respectively.
>
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
