Fix Databricks get_tables non-table response handling

[ianu82]

Summary

• add a guard in Databricks get_tables() to return early for non-table responses or missing dataframes
• only uppercase/rename columns when a table dataframe is present
• add a regression unit test ensuring error responses pass through unchanged

Why

get_tables() previously assumed result.data_frame always existed. On error/non-table responses this triggered secondary exceptions and masked the original failure.

Validation

• python3 -m py_compile mindsdb/integrations/handlers/databricks_handler/databricks_handler.py tests/unit/handlers/test_databricks.py
• pytest -q tests/unit/handlers/test_databricks.py -k get_tables_returns_non_table_response_without_transform (fails in this environment due to missing optional dependency appdirs)

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[ianu82]

review