[
https://issues.apache.org/jira/browse/FLINK-18449?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17151709#comment-17151709
]
Jingsong Lee commented on FLINK-18449:
--------------------------------------
If we respect DataStream API, In {{FlinkKafkaConsumer}} , the constructor is
{{List<String> topics, Pattern subscriptionPattern}} . I think it is good. So I
second Jark's proposal.
Do we really need a "topic-type"? If there is a third way to configure kafka
topic, I think we need have a consideration. Otherwise, we don't need.
> Make topic discovery and partition discovery configurable for
> FlinkKafkaConsumer in Table API
> ---------------------------------------------------------------------------------------------
>
> Key: FLINK-18449
> URL: https://issues.apache.org/jira/browse/FLINK-18449
> Project: Flink
> Issue Type: Improvement
> Components: Connectors / Kafka, Table SQL / API
> Affects Versions: 1.11.0
> Reporter: Shengkai Fang
> Assignee: Shengkai Fang
> Priority: Major
> Fix For: 1.12.0
>
>
> In streaming api, we can use regex to find topic and enable partiton
> discovery by setting non-negative value for
> `{{flink.partition-discovery.interval-millis}}`. However, it's not work in
> table api. I think we can add options such as 'topic-regex' and
> '{{partition-discovery.interval-millis}}' in WITH block for users.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
