[jira] [Created] (HIVE-10363) Provide a way to normalize the legacy partition column values

["Chaoyu Tang (JIRA)" <ji...@apache.org>]

Chaoyu Tang created HIVE-10363:
----------------------------------

             Summary: Provide a way to normalize the legacy partition column values
                 Key: HIVE-10363
                 URL: https://issues.apache.org/jira/browse/HIVE-10363
             Project: Hive
          Issue Type: Improvement
          Components: Types
            Reporter: Chaoyu Tang
            Assignee: Chaoyu Tang


We have seen a lot issues which were caused by the non-normalized partition column values, such as HIVE-10210, HIVE-6052 etc. Besides type checking, converting and normalizing the partition column values in insert/alter partition operations (see HIVE-10307, HIVE-10362), we need provide an easy way for users to normalize their legacy partition column data. HIVE-5700 has attempted to do at metastore sql level, but given that many flavors of backend databases and different versions, it is quite hard to achieve that and it is also error prone. The sql portion of change in HIVE-5700 has been reverted by HIVE-9445/HIVE-9509.
Currently alter table .. partition ... rename could be used to normalize the partition column for each partition, but I am thinking any other more convenient and better way to do that.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)