imonteroq commented on PR #59299: URL: https://github.com/apache/spark/pull/59299#issuecomment-6192573710
> ### Connectors vet the unpositioned AddColumn but receive the positioned one > > **Question:** Is failing the intended behavior when a connector cannot honor positions? One alternative is to re-check each positioned change with `supportsColumnChange` and fall back to a plain append when it is rejected. Later `AFTER` anchors stay valid because the column is still added. Either way, should the `SupportsSchemaEvolution` Javadoc say that an accepted `AddColumn` may carry a position assigned after the check? Keeping the failure. The flag is an explicit request for ordering, so silently appending would hide that the sink can't honour it. Documented in `SupportsSchemaEvolution` that accepted `AddColumn`s may gain a `ColumnPosition` before `alterTable`; connectors that can't position can reject there, matching `ALTER TABLE ... ADD COLUMN AFTER` today. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
