[
https://issues.apache.org/jira/browse/HIVE-13050?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15212818#comment-15212818
]
Lenni Kuff commented on HIVE-13050:
-----------------------------------
Do we even want to fix this? It seems much simpler and easier to understand the
behavior if we do not de-dupe locations. There also may be existing users who
expect the current behavior so it would be an incompatible change.
> The row count is not correct after changing partition location to point to
> another partition location
> -----------------------------------------------------------------------------------------------------
>
> Key: HIVE-13050
> URL: https://issues.apache.org/jira/browse/HIVE-13050
> Project: Hive
> Issue Type: Improvement
> Components: Query Processor
> Affects Versions: 2.1.0
> Reporter: Aihua Xu
> Assignee: Aihua Xu
>
> {noformat}
> CREATE TABLE test (s STRING) PARTITIONED BY (p SMALLINT) location
> 'data/test';
> INSERT INTO test PARTITION (`p`=1) VALUES ("v1");
> INSERT INTO test PARTITION (`p`=2) VALUES ("v2");
> ALTER TABLE test PARTITION (`p`=2) SET LOCATION '/data/test/p=1';
> {noformat}
> {{select * from test;}} shows 2 rows while {{SELECT count(*) FROM test;}}
> shows 1.
> That is inconsistent.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
