RLS managed ‚Deanonymization‘

Hi Raketenrudi 

RLS is applied at the semantic layer, so it cannot be overridden by measures. Calculated columns also do not change based on who is logged in, so that approach will not work. You will need a table that is not affected by RLS and use measures to show Ben’s name while classifying the rest. I have a YouTube tutorial that explains how to anonymize other users while keeping the name of the currently logged-in user visible. https://youtu.be/qL_p6OLDyQI 

Hi @Raketenrudi, as danextian  and AWD , already mentioned, RLS only applies a row‑level filtering context. However, if you were under pressure and needed an urgent workaround, we can try follow these steps:

1. Create the two tables without relationships, including an email column for RLS based on USERPRINCIPALNAME()

2. Use the TID field from MAIN_ as the basis of the visual
Create a table or matrix visual and place:
Rows: MAIN_[TID]
Values: the measures you will create in the next step (for example, Display Name, SUM(MAIN_[Val1]))
If needed, you may later hide the TID row header so that only names or “Classified” appear.

3. Create the required DAX measures: My IDs, My Name, and Display Name
The following measures support the RLS logic where:
Regular users see only their own TID in NAMES_, and therefore their own name is revealed.
Users without RLS (for example administrators) see multiple or all TIDs in NAMES_, and therefore all values remain anonymised.

3.1. Measure: My IDs
This returns the user’s own TID if RLS restricts them to one row. If more than one row is visible (meaning the user has unrestricted access), it returns BLANK, ensuring no de-anonymisation occurs.

My IDs :=
VAR VisibleTIDs =
CALCULATETABLE (
VALUES ( NAMES_[TID] )
)
VAR CountTIDs =
COUNTROWS ( VisibleTIDs )
RETURN
IF (
CountTIDs = 1,
MAX ( NAMES_[TID] ),
BLANK ()
)
3.2. Measure: My Name
Returns the real name of the current user only when the RLS context restricts NAMES_ to exactly one row. Otherwise returns BLANK.

My Name =
VAR VisibleName =
CALCULATE (
MAX ( NAMES_[Name] ),
ALL ( NAMES_ ) -- respeita RLS, mas remove outros filtros
)
RETURN
VisibleName
 
3.3. Measure: Display Name
This determines what is shown on each row of the visual. If the row belongs to the current user, it reveals their name; otherwise it displays “Classified”.

Display Name =
VAR VisibleName = [My TIDs]
var SelectedID = SELECTEDVALUE(Main_[TID])
var _Total = HASONEVALUE(Main_[TID])

RETURN
SWITCH(TRUE(),
_Total = FALSE(), BLANK(),
VisibleName = SelectedID,[My Name],
"Classified")
 
4. Create the RLS rule based on the NAMES_ table
Go to Model view → Manage Roles
Create a role, for example, RLS_Users
On the NAMES_ table, apply the following filter:
 
 
DISCLAIMER: While I wrote a draft of this answer, I used Copilot to create a longer, more detailed step-by-step description to make it easier to apply.

The file is attached.

With RLS, I do not think you can achieve what you are looking for within a single measure pointed at the same table with RLS applied. RLS will remove the other rows from 'Ben' meaning Ben's total will only be 8.

I am also not sure if you could display 'classified' per row like that for Ben to see row level context without seeing names (short of creating multiple agregated tables which is probably not feesable)

To allow Ben (and other users) to see the total of 15 would be possible with an agregated table, however it would look something like this (i.e. missing the 'classified' rows:

1) Apply RLS so users only ever see their own rows

2) Precompute global averages in:

• Power Query
• SQL
• Dataflow

3) Store them in a separate table not affected by RLS

4) Use that table for the “total” values in visuals by HASONEVALUE() expression.