[
https://issues.apache.org/jira/browse/TAJO-1109?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14191560#comment-14191560
]
ASF GitHub Bot commented on TAJO-1109:
--------------------------------------
Github user ykrips commented on the pull request:
https://github.com/apache/tajo/pull/216#issuecomment-61229439
Well, I did not remember who stated this technique. I feel that it may
comes for a need for separation of roles for DBA and programmer, or it may
comes for easier maintenance. At first, I would liked to add codes for sql
separations on catalog stores, but I added some codes for altering and deleting
databse objects because I prefer to automation on database schema modification.
Also, I found a similar document in a below link.
* http://www.javapractices.com/topic/TopicAction.do?Id=105
> Separate SQL Statements from Catalog Stores
> -------------------------------------------
>
> Key: TAJO-1109
> URL: https://issues.apache.org/jira/browse/TAJO-1109
> Project: Tajo
> Issue Type: Improvement
> Components: catalog
> Reporter: Jihun Kang
> Assignee: Jihun Kang
> Priority: Minor
> Fix For: 0.9.1
>
> Attachments: DBMSSchemaDefinition.xsd, derby.xml
>
>
> When developing the additional catalog stores for another database systems,
> it is needed to add additional sql statements such as triggers and indexes.
> These sql statements could increase the code size of catalog store, and when
> database schema has changed, it may affect the catalog store source codes. I
> feel that it is needed to separate the sql statements from the java source
> codes.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
