[
https://issues.apache.org/jira/browse/FLINK-41366?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated FLINK-41366:
-----------------------------------
Labels: pull-request-available (was: )
> Support disabling automatic column addition in the Fluss pipeline sink
> ----------------------------------------------------------------------
>
> Key: FLINK-41366
> URL: https://issues.apache.org/jira/browse/FLINK-41366
> Project: Flink
> Issue Type: Improvement
> Components: Flink CDC
> Reporter: Xiaobing Fang
> Priority: Major
> Labels: pull-request-available
>
> h3. Description
> The Fluss pipeline sink currently issues ADD COLUMN DDL when processing
> supported column-addition events. There is no option to disable that DDL
> while continuing to update the pipeline and sink schemas. Setting
> \{{schema.change.behavior: IGNORE}} drops the events and does not provide
> this behavior.
> h3. Proposed behavior
> Add a Boolean sink option \{{schema.evolution.auto-add-columns}}, defaulting
> to \{{true}} to preserve current behavior.
> When set to \{{false}}:
> * Validate the target schema without issuing ADD COLUMN DDL.
> * Validate during per-table initialization, including after recovery, and on
> each forwarded ADD event. Fail if required columns are missing or have
> incompatible types or nullability.
> * Refresh internal schemas and column mappings after successful validation so
> subsequent records can use the added columns.
> Keep \{{existing-table.schema-expansion.mode}}, table creation and other DDL
> behavior unchanged. ADD events must still reach the sink.
> Related: [FLINK-40349|https://issues.apache.org/jira/browse/FLINK-40349]
> proposes schema compatibility modes. This issue adds control over ADD COLUMN
> execution.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
