Niklaus Xiao created HIVE-17080:
-----------------------------------
Summary: Overwrite does not work when multi insert into same table
different partition
Key: HIVE-17080
URL: https://issues.apache.org/jira/browse/HIVE-17080
Project: Hive
Issue Type: Bug
Components: SQL
Affects Versions: 2.3.0
Reporter: Niklaus Xiao
Priority: Critical
Repo:
{code}
create table foo(k string, v string) partitioned by (dt string) stored as orc;
insert into table foo partition(dt='00') values('k1', 'v1');
insert into table foo partition(dt='01') values('k2', 'v2');
{code}
Run query:
{code}
from (
select * from foo
) t
insert overwrite table foo partition(dt='00') select t.k, t.v where t.k = 'k1'
insert into table foo partition(dt='01') select t.k, t.v where t.k = 'k2';
{code}
Note that is overwrite for partition dt='00'
expect result
{code}
0: jdbc:hive2://localhost:10000/default> select * from foo;
+--------+--------+---------+
| foo.k | foo.v | foo.dt |
+--------+--------+---------+
| k1 | v1 | 00 |
| k2 | v2 | 01 |
| k2 | v2 | 01 |
+--------+--------+---------+
4 rows selected (0.125 seconds)
{code}
Actual result:
{code}
0: jdbc:hive2://localhost:10000/default> select * from foo;
+--------+--------+---------+
| foo.k | foo.v | foo.dt |
+--------+--------+---------+
| k1 | v1 | 00 |
| k1 | v1 | 00 |
| k2 | v2 | 01 |
| k2 | v2 | 01 |
+--------+--------+---------+
4 rows selected (0.125 seconds)
{code}
Explain result says MoveTask for partition dt = '0' repalce is false:
{code}
| Stage: Stage-0 |
| Move Operator |
| tables: |
| partition: |
| dt 00 |
| replace: false |
| table: |
| input format: org.apache.hadoop.hive.ql.io.orc.OrcInputFormat |
| output format: org.apache.hadoop.hive.ql.io.orc.OrcOutputFormat
|
| serde: org.apache.hadoop.hive.ql.io.orc.OrcSerde |
| name: default.foo
{code}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
