This issue still exists unfortunately.
I'm hoping for the same solution as the poster had (i.e. that is goes away by itself), but it appears to me this is what Microsoft claims should not be possible, the result of a corrupt file. The dataset consists of many separate files (XML / CSV / XLS(X)) from a couple of sources, all cleaned and combined into one specific table. This is modelled together with data from a SQL Server instance.
I've had problems with this same report before where one critical error in one of the files, all transformed with the help of the auto-generated function, based on the transformations of an example report, would delete EVERY source in the query editor that is located directly or in a subfolder under the same top folder as where that auto-generated function would reside, followed by a long list of suddenly "Missing" tables when trying to load the data.
Other strange behaviour of this same file is that:
- after changing queries in the query editor and refreshing and republishing the report, the online version, would initially show the changes I made including the refreshed data, yet after clicking anything in the report, it would change itself (online) to revert to the old version. Not even waiting a while or refreshing all visuals will change this. A manual refresh also didn't solve the problem, however, the scheduled refresh two days later did.
- Also, reopening the file in PBI Desktop, it would ask me again to load or discard the changes made in the query editor (which I already loaded before refreshing and republishing).
- then, the reason I'm writing this here is that even though the scheduled refresh now does include the changes and new data, it is not possible to right click the data in the tables (which worked like a charm before and nothing changed there) to drill-through to different pages in the same report. The menu simply doesn't show.
- and finally, I see that a shape map doesn't fill anymore, which it always did before.
My best guess is that this is related to the amount of memory required for this type of data loading. I'll test with less data and see if this will solve the problem. I'm guessing the only real solution is probably to reorganize the flow on the front-end and just pull all the data from the SQL Server database.
Cheers,
Niels
p.s. just before I'm posting this, I clicked F5 in the browser (Edge) and now it works? It seems like a hickup, but it's definitely strange behaviour.
