Chris Kudelka created HIVE-8068:
-----------------------------------
Summary: Dynamic partition insert overwrite does not overwrite
files, but instead appends.
Key: HIVE-8068
URL: https://issues.apache.org/jira/browse/HIVE-8068
Project: Hive
Issue Type: Bug
Components: SQL
Affects Versions: 0.13.1
Environment: Centos 6.2; Amazon S3 as DFS
Reporter: Chris Kudelka
-- using a reference table `one_row` with contents:
|dummy_field|
dummy_value
-- create test table
create table if not exists test_table (
line string
) partitioned by (
my_part string
)
-- run first time
insert overwrite table test_table partition (my_part) select 'a', 'partVal'
from one_row;
Partition test_db.test_table{my_part=partVal} stats: [numFiles=1, numRows=1,
totalSize=10, rawDataSize=1]
-- run again
insert overwrite table test_table partition (my_part) select 'a', 'partVal'
from one_row;
Partition test_db.test_table{my_part=partVal} stats: [numFiles=2, numRows=1,
totalSize=20, rawDataSize=1]
select * from test_table where my_part = 'partVal';
Expected result:
1 row
Actual result:
2 rows
----
-- compare to non-dynamic partition insert, which overwrites as expected
----
-- drop table and recreate with sme definition
insert overwrite table test_table partition (my_part='partVal') select 'a' from
one_row;
Partition test_db.test_table{my_part=partVal} stats: [numFiles=1, numRows=1,
totalSize=10, rawDataSize=1]
-- run again
insert overwrite table test_table partition (my_part='partVal') select 'a' from
one_row;
Partition test_db.test_table{my_part=partVal} stats: [numFiles=1, numRows=1,
totalSize=10, rawDataSize=1]
select * from test_table where my_part = 'partVal';
Expected result:
1 row
Actual result:
1 row
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)