Users get different results in report (consumed via App)

[Vegard1985]

Hi,

We have faced some seemingly strange results sometimes in our reports for different users.

It seems to be a recurring issue. I don't think our setup is strange:

1. A dataset workspace where RLS is applied (underlying datasets)

2. Report workspace where we develop reports, publish and update the app, using dataset from 1)

3. End users only have access to the data according to RLS in the underlying dataset, and only consumes through the report workspace App (with different audiences).

4. Changes to reports are done in Desktop, uploaded and replacing existing reports, then updating app.

Some users experience weird kind of "transient" issues, at least I'm not able to understand or reproduce based on the filters that they have applied. I don't know how many times I've been asked to have a look - the filters looks fine and correct, but numbers are sometimes "misplaced" or yeah, strange things can happen, numbers can be quite off.

According to this topic, there was some talk about users filtering the reports (of course), then when the report is updated/replaced, the filters was bugging out a bit, filtering the data even though the actual slicers were not applying any filters, and that they were able to correct this with using the "Reset filter" button.

"Different users see different results in same Power BI report published to the service" 

This is something that often helps also here (unfortunately!).

If I ask the user to click on the "Reset filter" and do a page refresh, this often clears out the issues. Ok??

But this of course very painful to try learn an entire organization. If they can't trust the data that they are looking at.. etc..

It just doesn't sit well with management that the reports are not 100% reliable.

Anyone else having some strange transient issues with different users getting different results?

This is very annoying and I don't seem to find any good answers.

Br

Vegard

[lbendlin]

Unfortunately this is working as designed.  Try to bundle your meta data changes more so that the disruptions for the users are not as frequent.  Always, always update the app immediately after making large meta data changes. (This is one of the very few reasons for considering deployment pipelines so that you can shield users better from  frequent changes caused by agile development.) 

[Vegard1985]

Hm, ok. Working as designed, but is it working as desired?

Yes, I'm trying as good as I can to bundle up the changes. And I always update the app immediately after I publish a report update (perhaps not necessary, but I just do it).

But this is quite irritating. There should be a possibility for the report creator to kind of clear the report state for end-users. I mean, what I would really like, is an option during "Update App" process, where you could say "clear and reset report filters for all users".

We do have the "Persistent filters" option on reports, what would happen if I set "Don't allow", then "Update App". Would that trigger a reset? Then I could go back, "Allowing" again and "Update App"?

[lbendlin]

You can experiment with that but in most cases it makes the UX worse.