[
https://issues.apache.org/jira/browse/FLINK-17826?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17340636#comment-17340636
]
Flavio Pompermaier edited comment on FLINK-17826 at 5/7/21, 7:26 AM:
---------------------------------------------------------------------
This feature was added in the old connector[1] and I think it becomes useless
if Flink allows to read only the necessary data: if I register the Flink
catalog a table for db1.A and 1 for db1.B is Flink able to optimize query
like "select * from table_A join table_B" without requiring to fetch all data
from A and B and then joining..? Wouldn't this be a waste of time?
If the problem is that it becomes a table while it should be a view (read-only)
maybe Flink could introduce the concept of VIEW. Maybe this could be a
reasonable trade-off for people coming from SQL world..
[1] https://issues.apache.org/jira/browse/FLINK-17361
was (Author: f.pompermaier):
This feature was added in the old connecto[1] and I think it becomes useless
if Flink allows to read only the necessary data (i.e. If I register the Flink
catalog a table for db1.A and 1 for db1.B is Flink able to optimize query
like "select * from table_A join table_B" without requiring to fetch all data
from A and B and then joining..? Wouldn't this be a waste of time?
If the problem is that it becomes a table while it should be a view (read-only)
maybe Flink could introduce the concept of VIEW. Maybe this could be a
reasonable trade-off for people coming from SQL world..
[1] https://issues.apache.org/jira/browse/FLINK-17361
> Add missing custom query support on new jdbc connector
> ------------------------------------------------------
>
> Key: FLINK-17826
> URL: https://issues.apache.org/jira/browse/FLINK-17826
> Project: Flink
> Issue Type: Bug
> Components: Connectors / JDBC
> Reporter: Jark Wu
> Assignee: Leonard Xu
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.14.0
>
>
> In FLINK-17361, we added custom query on JDBC tables, but missing to add the
> same ability on new jdbc connector (i.e.
> {{JdbcDynamicTableSourceSinkFactory}}).
> In the new jdbc connector, maybe we should call it {{scan.query}} to keep
> consistent with other scan options, besides we need to make {{"table-name"}}
> optional, but add validation that "table-name" and "scan.query" shouldn't
> both be empty, and "table-name" must not be empty when used as sink.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
