You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@flink.apache.org by "Leonard Xu (Jira)" <ji...@apache.org> on 2024/11/21 07:26:00 UTC

[jira] [Assigned] (FLINK-36763) Support schema inference and evolution with single-table-mutliple-partition sources

     [ https://issues.apache.org/jira/browse/FLINK-36763?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Leonard Xu reassigned FLINK-36763:
----------------------------------

    Assignee: yux

> Support schema inference and evolution with single-table-mutliple-partition sources
> -----------------------------------------------------------------------------------
>
>                 Key: FLINK-36763
>                 URL: https://issues.apache.org/jira/browse/FLINK-36763
>             Project: Flink
>          Issue Type: Bug
>          Components: Flink CDC
>            Reporter: yux
>            Assignee: yux
>            Priority: Major
>
> Current schema evolution implementation implicitly assumes that there can't be two partitions with the same TableID being consumed on different subTasks. This is not always true for some weak-structured sources like Kafka and MongoDB, where table schemas are not centralized and might evolve independently in various parallelized tasks.
> It's very unlikely to be a trivial change, and might involves modifying most part of pipeline implementation.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)