SEZ9 commented on issue #12601: URL: https://github.com/apache/seatunnel/issues/12601#issuecomment-6537566908
@fatmanverse thanks for re-running the issue's steps on dev e2957d2 and for narrowing down where the bare `typname` is kept on the pgoutput RELATION path. Please go ahead. Resolving the spelling with `format_type(oid, NULL)` from the OID that `TypeRegistry.get(int oid)` already resolves sounds reasonable to me: it gives quoting and schema qualification in one place, matches how the initial `CREATE TABLE` already spells enum types, and leaves `PostgresType#getName()` and the name-keyed cache untouched. Failing the job with table, column and OID when the OID cannot be resolved is also what I'd expect. Your comment was cut off at the test section — could you post the rest of the test plan? Ideally the e2e case covers the rows in your table, including the type in a schema that is not on the `search_path`. @SEPURI-SAI-KRISHNA the `TypeRegistry` findings you listed (namespace joined but not selected, cache keyed on bare `typname`, `get(String)` dropping the schema prefix) are not on the OID path this fix takes, but they still apply to the name-keyed lookups. A separate issue for the multi-schema collision, as you offered, would be welcome. <!-- streview-comment:2151 --> -- 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]
