MetricSign
EN|NLRequest Access
Medium severityconfiguration

Power BI Refresh Error:
TAG_NOT_FOUND

What does this error mean?

An ALTER or CREATE statement tried to apply or reference a Snowflake object tag that does not exist in the specified schema, causing the tag assignment to fail.

Common causes

  • 1The tag was dropped and a policy or automation script still references it
  • 2The tag name is misspelled or the three-part qualified name points to the wrong database or schema
  • 3The tag was created in a dev schema and production pipelines reference it without a fully qualified name
  • 4RBAC or a schema cloning operation did not include the tag object

How to fix it

  1. 1Step 1: Run SHOW TAGS IN SCHEMA <database>.<schema> to list available tags.
  2. 2Step 2: If the tag is missing, recreate it: CREATE TAG <db>.<schema>.<tag_name> ALLOWED_VALUES 'value1', 'value2'.
  3. 3Step 3: Update all ALTER TABLE / ALTER COLUMN statements that apply the tag to use the fully qualified three-part name.
  4. 4Step 4: For tag-based masking policies, verify the tag exists before applying the policy binding.
  5. 5Step 5: Audit downstream governance automations (tag propagation rules, masking policy bindings) to remove references to deleted tags.

Frequently asked questions

Can I apply a tag to a Snowflake column without a schema_level CREATE TAG first?

No. Tags must be created as schema-level objects before they can be applied to tables, columns, or warehouses. You cannot create ad hoc tag strings at assignment time.

What happens to tag-based masking policies if the underlying tag is dropped?

Dropping the tag does not automatically remove the masking policy binding, but it prevents new tag assignments. Existing bindings that reference the dropped tag will show a policy error in Snowsight until the tag is recreated or the binding is removed.

Other configuration errors