[
https://issues.apache.org/jira/browse/TAJO-1253?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14249444#comment-14249444
]
Hyunsik Choi commented on TAJO-1253:
------------------------------------
Designing this kind of APIs requires careful thinkings. Basically, we need to
have the following design considerations:
* API Usability - should be easy to use and intuitive.
* Flexibility - allows users to put various predicates.
* Efficiency - each catalog store implementation should make good use of
indexes against given predicates.
> Catalog API should be queryable
> -------------------------------
>
> Key: TAJO-1253
> URL: https://issues.apache.org/jira/browse/TAJO-1253
> Project: Tajo
> Issue Type: New Feature
> Components: catalog
> Reporter: Hyunsik Choi
>
> TAJO-1176 is trying to add queryable catalog information. But, currently
> Catalog API is not quryable. In other words, Catalog retrieves all tables,
> and all tablespaces, and all functions. It is not scalable.
> It would be great if Catalog API takes some conditions like {{table_name LIKE
> 'T%'}} or {{table_name = 'orders'}}. This approach can be adopted to
> functions, tables, columns, tablespaces, and partitions.
> We already use RDBMS as catalog persistent storages. All RDBMSs provides
> abundant features, including indexing features. They allow very efficient
> implementations
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
