Hello @andreapeterson 

Yes, your understanding of Databricks tag behavior is correct. In Databricks Unity Catalog, tags follow a hierarchical inheritance pattern:

Downward inheritance: Tags applied at higher levels (catalog → schema → table) become available in the dropdown menus for lower levels. This allows you to easily apply organizational or governance tags consistently across related assets.

Column-level isolation: Tags created at the column level don't automatically propagate to other columns' dropdown menus.
This is by design to prevent dropdown clutter from highly specific, single-use column tags.
This behavior makes sense in Databricks' governance model because:
- Catalog-level tags are typically for broad organizational categories (like "PII", "Finance", "Marketing")
- Schema-level tags might represent departmental or project-specific classifications
- Table-level tags could indicate data types or processing stages
- Column-level tags are usually very specific descriptors that don't need to be reused elsewhere

Workaround for frequently used column tags: If you find yourself repeatedly typing the same column-level tag, you can:
- Apply it at a higher level first (table/schema) to make it available in dropdowns
- Use it on columns as needed
- Optionally remove it from the higher level later if it's not actually applicable there

This hierarchical approach aligns with Databricks' emphasis on data governance and helps maintain clean,
organized tagging without overwhelming users with irrelevant options in their dropdown menus.