[
https://issues.apache.org/jira/browse/FLINK-10232?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16841511#comment-16841511
]
Rong Rong commented on FLINK-10232:
-----------------------------------
Thanks Danny for bringing this up.
I think we can definitely collaborate on this together. I think based on the
[email
thread|http://apache-flink-mailing-list-archive.1008284.n3.nabble.com/DISCUSS-Flink-SQL-DDL-Design-tp25006.html]
we've pretty much agreeing with going forward with the MVP design on the
{{CREATE TABLE}} functionality first.
I would follow up with [~suez1224] offline to see what his opinion on the
status on FLINK-6962.
I think it is appropriate to follow up on this is multiple angle.
1. [~danny0405] can you share more details on the *flink-sql-parser* you
mentioned and how you envision to first support DDL on the batch side first?
AFAIK currently the flink-table-planner is supporting both the dataset and
datastream side at the same time and I think flink-table-planner-blink is a
single module. I think it would be nice to layout what's the
structure/dependency vision for discussion of backward compatibility; migration
and plans.
2. I will follow up with the discussion thread shortly to bring back some of
the points regarding the post MVP discussion which has some discrepancies
between the design doc, such as attribute / source,sink.
3. I think we can also follow up with some of the orthogonal implementations
for the rest of the discussion, such as: {{CREATE FUNCTION}}.
Move this piece by piece is a good idea IMO. what do you guys think?
[~danny0405][~ykt836][~suez1224][~twalthr]
> Add a SQL DDL
> -------------
>
> Key: FLINK-10232
> URL: https://issues.apache.org/jira/browse/FLINK-10232
> Project: Flink
> Issue Type: New Feature
> Components: Table SQL / API
> Reporter: Timo Walther
> Assignee: Danny Chan
> Priority: Major
>
> This is an umbrella issue for all efforts related to supporting a SQL Data
> Definition Language (DDL) in Flink's Table & SQL API.
> Such a DDL includes creating, deleting, replacing:
> - tables
> - views
> - functions
> - types
> - libraries
> - catalogs
> If possible, the parsing/validating/logical part should be done using
> Calcite. Related issues are CALCITE-707, CALCITE-2045, CALCITE-2046,
> CALCITE-2214, and others.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)