[
https://issues.apache.org/jira/browse/HIVE-15118?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15743236#comment-15743236
]
Naveen Gangam commented on HIVE-15118:
--------------------------------------
[~aihuaxu] Functionally, the second version of the patch looks fine. Just a
couple of nits:
1) the {{UPDATE "APP".VERSION SET SCHEMA_VERSION=}} command should be the last
command in this file for a couple of reasons a) for consistency with other
upgrade schema files and more importantly 2) the version in this table is
pretty heavily relied upon to determine the schema version. So the upgrade will
not be considered complete until the version is set. If for some reason, the
upgrade script terminates abruptly before this version is set, the version
would not be set and would be and indication that something went wrong during
upgrade. With the current patch, if the script ends right after the version is
set, the drop table command will not be processed but the schema version would
indicate that it succeeded.
2) Can you move the {{DROP TABLE}} command to a separate file, say
037-HIVE-15118.derby.sql and then just call RUN on this file from the upgrade
script? Thanks
> Remove unused 'COLUMNS' table from derby schema
> -----------------------------------------------
>
> Key: HIVE-15118
> URL: https://issues.apache.org/jira/browse/HIVE-15118
> Project: Hive
> Issue Type: Sub-task
> Components: Database/Schema
> Affects Versions: 2.2.0
> Reporter: Aihua Xu
> Assignee: Aihua Xu
> Priority: Minor
> Attachments: HIVE-15118.1.patch, HIVE-15118.2.patch
>
>
> COLUMNS table is unused any more. Other databases already removed it. Remove
> from derby as well.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
