The ultimate Microsoft Fabric, Power BI, Azure AI, and SQL learning event: Join us in Stockholm, September 24-27, 2024.
Save €200 with code MSCUST on top of early bird pricing!
Find everything you need to get certified on Fabric—skills challenges, live sessions, exam prep, role guidance, and more. Get started
Hello Community,
I have two tables, Table A (state, population) Table B (username, state).
Is there a way to setup RLS using username() to reference Table B(username) column to return the value of Table B(state) and join the data to Table A(state)?
On a separate but related question - Is there a way to assign a Parameter using UserName()?
Your help is much appreciated.
Solved! Go to Solution.
Hi there, you can most certainly do this.
I would suggest reading my blog post, which just gives some explanations around when deploying to the Power BI Service and what to look out for.
And then here is a great blog post by Kasper De Jonge, in which he gives a working example how to get RLS working. And also provides a sample PBIX file.
http://www.kasperonbi.com/power-bi-desktop-dynamic-security-cheat-sheet/
Hi there, you can most certainly do this.
I would suggest reading my blog post, which just gives some explanations around when deploying to the Power BI Service and what to look out for.
And then here is a great blog post by Kasper De Jonge, in which he gives a working example how to get RLS working. And also provides a sample PBIX file.
http://www.kasperonbi.com/power-bi-desktop-dynamic-security-cheat-sheet/
BTW, how come there's not a way to assign username() function within a parameter? If there's a way, I could restrict data at the import layer instead of at filtering layer to improve performance.
Any idea?
Hi there,
That is a by design feature and in order for RLS to work correctly, you always have to have the entire dataset.
This ensures that if you have a user who needs access to everything this can be achieved.
This also ensures that dynamic RLS can be achieved without having to re-create not only the charts but also the dataset.
If I understand you correctly you would want to only bring in the data for the specific user?
If that is the case you would then have to have a specific dataset per user, which would increase the size of each individual model.
By keeping all the data in one model and leveraging RLS ensures that the data is complete and consistent across the different users.
You addressed my concern perfectly. Thank you once again!
Hi
why i don't have (on the join) the option for “Apply security filter in both directions”...?
attach picture
User | Count |
---|---|
109 | |
82 | |
74 | |
53 | |
49 |
User | Count |
---|---|
133 | |
97 | |
94 | |
71 | |
69 |