[
https://issues.apache.org/jira/browse/HIVE-4891?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13773683#comment-13773683
]
Harish Butani commented on HIVE-4891:
-------------------------------------
I am not able to reproduce. Tried the scenario below, get correct results.
Can you provide us details on how to reproduce.
{noformat}
CREATE TABLE part(
p_partkey INT,
p_name STRING,
p_mfgr STRING,
p_brand STRING,
p_type STRING,
p_size INT,
p_container STRING,
p_retailprice DOUBLE,
p_comment STRING
)
;
LOAD DATA LOCAL INPATH '.../hive/data/files/part_tiny.txt' overwrite into table
part;
CREATE TABLE part_partitioned(
p_partkey INT,
p_name STRING,
p_mfgr STRING,
p_brand STRING,
p_type STRING,
p_size INT,
p_container STRING,
p_retailprice DOUBLE,
p_comment STRING
)
PARTITIONED BY (ds string);
alter table part_partitioned add partition(ds='2010');
alter table part_partitioned partition(ds='2010') set fileformat sequencefile;
-- have to do this, otherwise sets partition's format to textinputformat
alter table part_partitioned set fileformat sequencefile;
INSERT OVERWRITE TABLE part_partitioned PARTITION (ds='2010')
SELECT * from part;
alter table part_partitioned add partition(ds='2011');
alter table part_partitioned partition(ds='2011') set fileformat rcfile;
-- have to do this, otherwise sets partition's format to textinputformat
alter table part_partitioned set fileformat rcfile;
INSERT OVERWRITE TABLE part_partitioned PARTITION (ds='2011')
SELECT * from part;
-- tried these, give the right results
select distinct p_mfgr from part_partitioned where (ds='2010' or ds='2011') and
p_size < 10;
select distinct p_name from part_partitioned where (ds='2010' or ds='2011');
{noformat}
> Distinct includes duplicate records
> -----------------------------------
>
> Key: HIVE-4891
> URL: https://issues.apache.org/jira/browse/HIVE-4891
> Project: Hive
> Issue Type: Bug
> Components: File Formats, HiveServer2, Query Processor
> Affects Versions: 0.10.0
> Reporter: Fengdong Yu
> Priority: Blocker
> Fix For: 0.12.0
>
>
> I have two partitions, one is sequence file, another is RCFile, but they are
> the same data(only different file format).
> I have the following SQL:
> {code}
> select distinct uid from test where (dt ='20130718' or dt ='20130718_1') and
> cur_url like '%cq.aa.com%';
> {code}
> dt ='20130718' is sequence file,(default input format, which specified when
> create table)
>
> dt ='20130718_1' is RCFile.
> {code}
> ALTER TABLE test ADD IF NOT EXISTS PARTITION (dt='20130718_1') LOCATION
> '/user/test/test-data'
> ALTER TABLE test PARTITION(dt='20130718_1') SET FILEFORMAT RCFILE;
> {code}
> but there are duplicate recoreds in the result.
> If two partitions with the same input format, then there are no duplicate
> records.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
