azure databricks error

The error code INTERNAL_ERROR.UC_INTERNAL_ERROR (UC-AUDKE) indicates that the failure is occurring within the Unity Catalog layer while Genie is attempting to retrieve sample data. This is generally different from a typical SQL syntax, connectivity, or permission error.

A few targeted checks that can help narrow down the cause:

Validate the underlying objects directly

• Run a simple SELECT * FROM catalog.schema.table LIMIT 10 against the same table or view from the SQL Warehouse used by Genie. Confirm the query completes successfully and returns data.

Check whether the issue is object-specific

• If the Genie space references multiple tables/views, try retrieving sample data from a different object.
  • If only one object fails, review whether it contains complex view definitions, cross-catalog references, or recently modified metadata.

Verify Unity Catalog access

• Ensure the user or service principal used by Genie has:
  • USE CATALOG
  • USE SCHEMA`
  • SELECT
• on all referenced objects, including any tables accessed indirectly through views.

Review table and view health -

• Confirm that the table/view still exists and is accessible.
• If this is a view, test the view definition independently to ensure all underlying dependencies resolve successfully.

Check for metadata inconsistencies

• Run DESCRIBE TABLE or SHOW CREATE TABLE/VIEW on the affected object.
• Occasionally, Genie's profiling operations expose catalog metadata issues that are not immediately visible during normal querying.

What the results tell you:

• If normal SQL queries, DESCRIBE, and metadata operations all succeed, but "Get sample data" continues to fail, the issue is likely isolated to the Genie sample-data retrieval path rather than the underlying data itself.
• If direct queries or metadata commands also fail, focus on the specific catalog, schema, table, or view configuration that Genie is attempting to access.

Help make this community better for everyone: if this answer resolved your issue, please accept it or leave an upvote. If not, share more details in a comment so we can continue the discussion and find the right solution.