[
https://issues.apache.org/jira/browse/BEAM-2009?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15974117#comment-15974117
]
Jean-Baptiste Onofré commented on BEAM-2009:
--------------------------------------------
That's a discussion we already had and the point is: why using a source when
it's not possible to estimate the size and split.
Sink is useless compared to the {{Write/WriteFn}} already available. Do you
mean the {{Sink}} abstract class ?
> support JdbcIO as source/sink
> -----------------------------
>
> Key: BEAM-2009
> URL: https://issues.apache.org/jira/browse/BEAM-2009
> Project: Beam
> Issue Type: New Feature
> Components: dsl-sql
> Reporter: Xu Mingmin
> Assignee: Jean-Baptiste Onofré
>
> support JdbcIO in both source/sink part:
> 1. as source, JdbcIO read data from databases that supports JDBC such as
> Oracle/MySQL/Cassandra/...;
> It leads to a bounded pipeline;
> 2. as sink, JdbcIO can persistent data from both unbounded and bounded
> pipeline;
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
