Earn a 50% discount on the DP-600 certification exam by completing the Fabric 30 Days to Learn It challenge.
Hey Community,
Strange question.
In the desktop I have text slicer for a Direct Query that use M parameters. The slicer (green box) is updating correctly in the desktop. It is Date start - Date End: Duration text string.
When I publish to the service, and make a change using M inputs or recipe, it does not update the slicer. PS, the slicer is a further refinement on the results returned from SQL using the M parameters.
Interestingly enough, after playing around for way too long, I noticed that the same column in the slicer must also be in the matrix itself. The screen shot shows a date time : date time header, which is different than the Date start : Date End: Duration text string
The question is, is this is a bug (desktop behavior vs. service) or does this behavior make sense? Why does it work in desktop but but not the service...
I submotted a ticket to Microsoft as well but I don't usually get helpful responses.
Solved! Go to Solution.
They will update the list, but not the selection.
Alright, finally resolved this. I turned off the automatic aggregation training. I guess this result is cached. Using the pipeline and re-deploying the report also made the report work (temporarily for a day), but after the refresh, the aggregations resulted in caching, so this was not a long term solution.
The report works as exepcted now. The above is not the solution. The slicer (green box) will update based on the inputs selected.
@v-yetao1-msft is this a bug? Should slicers be cached? Should there be a feature to supress slicer caching?
A user had a similar issue where you helped here. I set capcity to defualt no caching. Still didn't work. Any ideas on how to add something like Rand to a text string? I am guessing this will prevent caching?
Also, oddly, in my Dev environment, the report works perfectly and not in Prod. I use Databricks FYI.
https://community.fabric.microsoft.com/t5/Desktop/How-do-I-prevent-caching-data-for-each-filter-used...
Hey,
Just wanted to update this. Had a call with Microsoft. After walking this through, the slicer magically worked. Can't tell if the new update fixed it but tried a few different things, including deleting and re-adding the slicer, abd clearing report cache. I blame the gremlins in the reort...
Please define what you mean by "Slicer update". Are you expecting the selection to change?
That's correct. Numeric parameters or a recipe change will update the selection list (green box). Basically, it's after all the M parameters have been applied, you can further refine the list.
I am debating running a query trace to see what is being sent back to SQL Server.
They will update the list, but not the selection.
Alright, finally resolved this. I turned off the automatic aggregation training. I guess this result is cached. Using the pipeline and re-deploying the report also made the report work (temporarily for a day), but after the refresh, the aggregations resulted in caching, so this was not a long term solution.
The report works as exepcted now. The above is not the solution. The slicer (green box) will update based on the inputs selected.
@v-yetao1-msft is this a bug? Should slicers be cached? Should there be a feature to supress slicer caching?