Get certified for free when you join Fabric Data Days 2026 and dive into Fabric, Power BI, SQL, AI, and other essential data skills.
Join nowData Days is here! Join us now for 60+ days of learning, challenges, and connection. Learn more
Hi all,
I'm using dataflow to preload data and then put said data into a larger dataset that also provides row-level security.
Starting out, I put the dataset/report and the dataflow into the same workspace and handed out the viewer permission on the workspace.
I noted that users with the viewer role can still access the sensitive data in dataflow using Power BI Desktop. That leads the RLS concept ad absurdum.
So I think a best practice is to setup separate workspace for the dataflows and then put the data model in a second workspace.
Can someone confirm or did I overlook some security option for dataflows?
Hi @trutz
You are correct, the best way is to keep the dataflows separate if you need let anyone access ( old days give access to the workspace ) or use certification capabilities for dataflows:
https://powerbi.microsoft.com/en-us/blog/announcing-new-certification-capabilities-for-dataflows/
Thanks @Mariusz for your reply.
Could you elaborate on your remark: "old days give access to workspace"?
What would your recommendation be given the fact that I have a model that others should be able to use in their reports.
Would you create an app instead of giving viewer access to the workspace?
Don't miss out on Data Days, June 15 through August 7. Learn Fabric, Power BI, SQL, AI and more.
Check out the May 2026 Power BI update to learn about new features.
| User | Count |
|---|---|
| 16 | |
| 11 | |
| 10 | |
| 8 | |
| 7 |
| User | Count |
|---|---|
| 41 | |
| 36 | |
| 36 | |
| 28 | |
| 20 |