Find everything you need to get certified on Fabric—skills challenges, live sessions, exam prep, role guidance, and more. Get started
In a tabular model for analysis services, the lowest level of a shared product dimension needs to be hidden from a a group of users if they select inventory measures, but they can see all product columns when selecting sales measures. There are about 20 product columns and only two need to be hidden for inventory measures. Is it possible to combine object level security with DAX to accomplish this?
I appreciate the suggestion of a shadow/hidden table, but please provide more details on how to implement. I'm also providing more specifics about what the model design and desired reporting output for two different user groups.
All Measures (calculated table for all explicit measures)
Product dimension (many attributes but limiting to 4 for this example)
Table relationships:
Example #1 for Sales user to see the first three Product fields:
Example #2 - Sales user should not see Material Number for Inventory Tons. Ideally the gray portion would be blank so the same report works for all users instead of completely hidden/access denied that would create an error on the report.
Example #3 - Non-Sales user can see all product levels:
Thanks for any suggestions!
use two copies of the Product Dimension table. Join the one with the visible materials column to the invoices and the one with then hidden materials column to the inventory
If inventory is joined to the version of the Product dimension with the material column hidden, wouldn't that hide material from all users? Most users should see inventory by material. It's only sales reps that need to have material hidden.
If you have sufficiently diverse audiences then you better create separate reports.
I'm using Tabular Editor for an Analysis Services cube that is accessed by the entire company in Excel for ad-hoc and some PBI reports. This requirement seems difficult to implement so was hoping some advanced DAX with OLS might make it possible.
I'm using Tabular Editor for an Analysis Services cube that is accessed by the entire company in Excel for ad-hoc and some PBI reports.
Would have been nice if you had led with that.
In that case I think you need two cubes.
My original post started with "In a tabular model for analysis services..."
Yes, it did. I misread it as referring to a Power BI dataset semantic model.
Use a shadow/hidden table instead of OLS.