FANNG1 commented on PR #10826:
URL: https://github.com/apache/gravitino/pull/10826#issuecomment-4301104362

   Thanks for the fix. Could you please confirm the exact trigger condition?
   
   From Spark analyzer order, `ResolveRelations` runs before 
`ResolveSQLOnFile`, so this issue seems to happen only when the current/default 
catalog is a Gravitino catalog. If the session stays on `spark_catalog` and 
only has `GravitinoSparkPlugin` registered, `SELECT * FROM parquet.`path`` 
should still fall through to Spark built-in SQL-on-file resolver and work 
normally.
   
   Could you verify both cases?
   
   1. Register `GravitinoSparkPlugin`, but do not `USE` a Gravitino catalog / 
do not set `spark.sql.defaultCatalog` to a Gravitino catalog. Does `SELECT * 
FROM parquet.`path`` work?
   2. Switch to a Gravitino catalog, or set `spark.sql.defaultCatalog` to one. 
Is this when the `MetadataObjects.of(TABLE, ...)` / `length of names must be 3` 
error appears?
   
   This distinction matters because the current guard treats any identifier 
with namespace `parquet`, `csv`, `json`, etc. as a DataSource shortcut. That 
may incorrectly hide a real Gravitino table under a schema named `parquet`, for 
example `gravitino_catalog.parquet.t`.


-- 
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]

Reply via email to