MaxGekk commented on pull request #30358:
URL: https://github.com/apache/spark/pull/30358#issuecomment-726577886


   The failed test is some kind of corner case. This guy
   
https://github.com/apache/spark/blob/fcf8aa59b5025dde9b4af36953146894659967e2/sql/core/src/test/scala/org/apache/spark/sql/connector/SupportsCatalogOptionsSuite.scala#L64
   assumes that if it passes `Array.empty` to `listTables()`, the function 
should return all tables.
   
   This works fine if `allNamespaces`:
   
https://github.com/apache/spark/blob/0bd7a3dfab41336dba2788a3d1fa3cf5b9f410d3/sql/catalyst/src/test/scala/org/apache/spark/sql/connector/InMemoryTableCatalog.scala#L127
   is not empty. But when `allNamespaces` is empty, there is no element with 
prefix `Array.empty`, and `namespaceExists` returns false:
   ```scala
     override def namespaceExists(namespace: Array[String]): Boolean = {
       allNamespaces.exists(_.startsWith(namespace))
     }
   ```
   
   As a possible solution, I would handle `Array.empty` in `listTables()`, and 
return all tables like:
   ```scala
     override def listTables(namespace: Array[String]): Array[Identifier] = {
       if (namespace.isEmpty || namespaceExists(namespace)) {
         super.listTables(namespace)
       } else {
         throw new NoSuchNamespaceException(namespace)
       }
     }
   ```


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
[email protected]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to