fatmanverse commented on issue #12601: URL: https://github.com/apache/seatunnel/issues/12601#issuecomment-6540700603
Thanks @SEZ9, I'll start on it. The test section should be complete in the comment above (it may be collapsed in the preview). The plan, in more detail: **Unit tests** - The type-name resolution for a RELATION column: a user-defined type takes the name returned for its OID, built-in types keep their current names, and an unresolved OID fails with the table, column and OID in the message. **E2E (`PostgresCDCIT`, ADD COLUMN with `schema-changes.enabled = true`)** - Pre-create in the shared source/sink database: schema `inv` (not on the `search_path`), enums `jobstatus_lc`, `"JobStatus"`, `"order"`, `inv.mood`, and the range type `"MyRange"`. - After the snapshot, add one column per type, which covers every row of the table above plus the range type, and insert a row that uses each new column. - Assert that each sink column's `format_type` equals the source column's (`jobstatus_lc`, `"JobStatus"`, `"order"`, `inv.mood`, `"MyRange"`), and that the sink rows equal the source rows. The sink URL in that case uses `stringtype=unspecified`, because the Jdbc sink binds enum values with `setString` (see #12588 and #13056). -- 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]
