The ultimate Fabric, Power BI, SQL, and AI community-led learning event. Save €200 with code FABCOMM.
Get registeredCompete to become Power BI Data Viz World Champion! First round ends August 18th. Get started.
using powerbi semantic models, how do i design a report to access it with a proxy account not pass through?
give you an example, I have a semantic model with rls setup and need to build two reports, one report will pass through user credentials and using rls. and another one will use proxy account to give the same access to any viewer. some of the users might not even have shared access to the semantic model. Once we have rls added in the model, I think we have to use them in order to run the report.
but if they do, they might be secured by rls instead of wide open in case the model will be shared. How can we keep them have the same access in the second report which means using proxy account
This can be easily implemented in powerbi/ssrs onpremise, but powerbi service only provide pass through option. If we do not have any role created, this is fine. but once we use roles, this becomes issues.
There will be two seperate reports. for example,
We have two separate financial reports built from Finance model.
Report 1: Budgetary Report (Manager View) This report focuses on cost centers and shows budgetary and actual spending data. It leverages Row-Level Security (RLS) to restrict data visibility based on a manager's assigned cost centers.
Report 2: Corporate Finance Statement Report (Full View) This report provides a high-level view of financial information for the entire corporation.
Once we add manager to RLS on the cost centers in the model, the first report works perfectly but the second one will only aggregate the info in their cost centers.
Because powerbi service only suppport pass through authentication we can not use proxy account for the second one like what we do in powerbi on-premise. The only option I can think of is to duplicate the same model with/without RLS added and use them for the reports seperately, in this way, it is not just resoures wasting, maintaining them to sync will be a big challege.
You can either use a shadow summary fact table that is not controlled by RLS or go with the approach you mentioned. I think the additional maintenance effort is manageable.
use separate reports and audiences.