Skip to content

Centralizing row and column access rules with Unity Catalog ABAC policies

Unity Catalog Attribute-Based Access Control (ABAC) policies let admins centrally define row filters and column masks using tags and conditions instead of hand-writing per-table SQL functions. Policies are attached at the catalog, schema, or table level and apply automatically to any matching object, reducing per-object governance overhead. This replaces manual row filter and column mask functions with a scalable, tag-driven approach.

1 · Learn the must-know

  • ABAC policies use governed tags (on columns, tables, or attributes) as conditions to decide when a row filter or column mask applies, so tagging data correctly is essential for policies to take effect.
  • Policies can be defined once at a catalog or schema level and automatically inherited by all matching tables/columns, avoiding the need to write a separate masking function for every table.
  • Column masking policies can redact, hash, or partially obscure sensitive column values based on the querying user's group membership or attributes, without altering the underlying stored data.
  • Row filter policies restrict which rows a user sees based on attribute matches (e.g., region or department tags), enforced transparently at query time regardless of how the data is accessed.
  • ABAC policies are managed via SQL (CREATE POLICY / governance UI) and require appropriate metastore-level privileges to create or attach; policy changes take effect without needing to rewrite views or tables.
  • Because enforcement happens centrally in Unity Catalog, ABAC policies apply consistently across all compute (SQL warehouses, clusters, notebooks) and access paths, unlike ad hoc view-based filtering.

2 · Check your understanding

Check this objectiveFree · always available

An ABAC policy applies a row filter so analysts only see rows where region matches their assigned region attribute. A user with no region attribute set queries the table. What happens?

Your objective map0 tried · 0 answered correctly · 33 untouched

What you have tried across Databricks DEA's objectives, not a readiness score.

Databricks Intelligence Platform6% of the exam0 of 2 tried
Data Ingestion and Loading21% of the exam0 of 7 tried
Data Transformation and Modeling22% of the exam0 of 7 tried
Working with Lakeflow Jobs16% of the exam0 of 4 tried
Implementing CI/CD10% of the exam0 of 4 tried
Troubleshooting, Monitoring, and Optimization10% of the exam0 of 5 tried
Governance and Security15% of the exam0 of 4 tried

3 · Keep going