Which of the following tools can be used to apply granular data access permissions in Fabric?

• A. OneLake data access roles
• B. DML
• C. Column-level security in the semantic model
• D. Row-level security in the semantic model

Answer: (A)

Granular data access permissions in Fabric are managed at the data layer through OneLake data access roles. OneLake acts as the unified data lake in Fabric, and data access roles let you define who can access which data assets and what they can do, at fine granularity—down to folders, datasets, tables, or files. This central policy is enforced across all tools and workloads, making it the best way to implement least-privilege access across the entire platform.

DML is just a data manipulation language; it does not provide a governance mechanism to restrict who can read or modify data. Column-level and row-level security in the semantic model restrict what appears inside a modeling layer, but they operate at the model level and depend on data being exposed through that model, rather than offering lake-wide, asset-level access control.