yuqi1129 opened a new issue, #14681: URL: https://github.com/apache/gravitino/issues/14681
### Version main branch (`bfa88fcbb41fe8d7fbb438778bb8cb8a9feeb28d`) ### Describe what's wrong Tag operations validate metadata objects through the catalog's normalizing dispatchers, but then use the original names to access the relation store. When the request spelling differs from the stored canonical name, an existing object is reported as missing. This affects listing/getting tags and associating/disassociating tags, including value assignments. Catalog-specific normalization, quoted names, and physical table-name resolution need to be respected consistently. ### Error message and/or stacktrace A regression test with an existing schema and table reproduces: ```text NoSuchMetadataObjectException: Failed to list tags for metadata object ... due to not found Caused by: NoSuchEntityException: No such schema entity: SCHEMA_FOR_TAG_TEST ``` ### How to reproduce 1. Use a catalog with case-insensitive schema/table names, such as Hive. 2. Create schema `analytics` and table `orders`; their canonical names are lowercase. 3. Associate a tag using `catalog.analytics.orders`. 4. List tags using the equivalent name `catalog.ANALYTICS.ORDERS`. The object existence check succeeds, but the relation lookup uses the uppercase names and fails. The same mismatch occurs when adding or removing tag assignments using noncanonical names. ### Additional context The failure is deterministic for the mismatched name; retrying the same lookup does not fix it. The metadata object identifier used by the relation store should follow the catalog's lookup rules, while existence checks must retain the original request to preserve quoted-name semantics. -- 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]
