Launch Deal · 50% off 1st month ·

4.6 — Row-Level Security and Sensitivity Labels

PL-300 Domain 4 · Skill 4.6

What You Need to Know

Row-level security (RLS)

RLS filters the rows a user can see, through roles defined in the semantic model. It's the only report-level way to make sure two people looking at the same report see different data.

1. Define roles in Power BI Desktop: Modeling > Manage roles > New. Pick a table and write a DAX filter expression that returns TRUE for the rows the role may see.

Type Example Use
Static [Region] = "West" A few fixed groups. One role per region
Dynamic [Email] = USERPRINCIPALNAME() One role for everyone. Each user sees their own rows, looked up from a table that maps users to what they may see

Dynamic RLS usually filters a user mapping table (e.g. SalesRep[Email], Region), and the filter flows to facts through relationships. USERPRINCIPALNAME() returns the sign-in name (UPN) in both Desktop and the service, so prefer it to USERNAME(), which returns DOMAIN\user in Desktop.

  • Filters propagate in the direction of relationships. If RLS on a dimension must flow through a bidirectional relationship, set Apply security filter in both directions on that relationship.
  • A user in several roles sees the union (all rows any of their roles allow).

2. Test in Desktop: Modeling > View as a role, or Other user plus a role to test dynamic RLS for a specific UPN. In the service, the model's Security page has Test as role.

3. Assign members in the service: open the semantic model > Security, select a role and add people or security groups (Entra ID security groups, mail-enabled security groups and distribution lists). Using groups means access changes are made in Entra ID, not in Power BI.

Who RLS applies to: users with the Viewer role, people who get access through an app or sharing, and users with only Build permission. Admins, Members and Contributors are not filtered. Roles defined in Desktop are only enforced after members are assigned in the service. A user with access to the report but no role sees no data once RLS exists.

For Direct Lake and DirectQuery models, RLS can also be defined in the source, but semantic-model RLS is what this exam focuses on.

Sensitivity labels

Microsoft Purview Information Protection sensitivity labels (for example Public, General, Confidential, Highly Confidential) classify Power BI and Fabric items.

  • Apply a label in Desktop (Home > Sensitivity) or in the service (item settings or the label in the header). Labels are created by the security team in Microsoft Purview, and the tenant setting Allow users to apply sensitivity labels must be on.
  • Protection follows exported data: exporting to Excel, PowerPoint or PDF, or downloading the .pbix, keeps the label and its encryption.
  • Inheritance: new reports and dashboards inherit the label from their semantic model (downstream inheritance), and items can inherit from data sources.
  • Admins can enforce default labels and mandatory labeling (an item can't be saved or published without a label).
  • A label doesn't restrict who can open the item inside Power BI. Access is still controlled by workspace roles, permissions and RLS.

Key Concept: RLS = DAX filter roles in the model, members assigned in the service, enforced only for Viewers, app and share users. Sensitivity labels classify content and protect what's exported, but don't replace permissions.

Exam Traps

  • Testing RLS as a workspace Member and concluding it doesn't work
  • Creating one static role per salesperson when a dynamic role with USERPRINCIPALNAME() is needed
  • Using USERNAME() and getting DOMAIN\user in Desktop
  • Forgetting to add members (or groups) to roles after publishing
  • Expecting a sensitivity label to stop a workspace Viewer from opening a report

Check your understanding

Practice Scenario

🔒 This practice scenario is part of Skillmarked Pro

The lessons stay free. Pro unlocks the other 22 PL-300 practice scenarios and the 50-question timed mock exam, plus everything for every other exam.

See plans

Launch Deal: first month €9.95, then €19.90/month. Enter code LAUNCH50 at checkout. Ends 1 December.

Cancel anytime. Or €199 a year, which saves 17%.

Not sure yet? Try 10 free mock exam questions.

Already Pro? Log in.