Forum Discussion

PRAVISH_SARAVAN's avatar
PRAVISH_SARAVAN
Frequent Visitor
2 years ago

Reports don't work post deploying from DEV to UAT using Deployment Pipeline

Visuals throw below error post deploying them to UAT but everything works fine in DEV.   Expression.Error: We couldn't fold the expression to the data source. Please try a simpler expression.. ;We ...
Need Help
PRAVISH_SARAVAN's avatar
PRAVISH_SARAVAN
Frequent Visitor
2 years ago

Thanks for the response. But the strange worry is,it's working smoothly in DEV workspace but why the issue raises in the TEST workspace alone post deployment.

If I publish directly from DESKTOP(Pointing DEV workspace) it's working but this issue occurs when I deploy to TEST workspace or any new workspace using PBI DEPLOYMENT PIPELINE.

 

So I'm confused on this part.

Also the calculcation used is normal CALCULATE(<AGGREGATE>,<FILTERs>) format

Brunner_BI's avatar
Brunner_BI
Icon for Impactful Individual rankImpactful Individual
2 years ago

PRAVISH_SARAVAN the first thing to understand here is that query folding happens in your Power Query, it has nothing to do with any calculations or DAX, measures etc.

 

Meaning that something in your backend/PQ breaks the DirectQuery.

 

Did you define some parameters there that might have changed when you pushed to UAT?

You should download whatever you have in UAT and analyze the issue in Power BI Desktop, because exactly that file cannot execute the direct query.

 

What I am guessing the problem is is that your UAT report was automatically binded to a different dataset. Do you have another pipeline for your datasets?

 

The error you are mentioning is quite weird though, only one visual is throwing this? Can you provide a screenshot?

  • PRAVISH_SARAVAN's avatar
    PRAVISH_SARAVAN
    Frequent Visitor
    2 years ago

    I do agree the Query folding happens in the Power Query Editor and we have not applied any query folding in our Power Query Editor.

     

    Yes we have connection details paramters defined in the Dataflows, pointing to these Dataflows the common Dataset(Semantic Model) is created.

     

    Inside dataset we perform basic transformation like join, append, remove or add columns.

     

    And no, as one workspace can be assigned only once, we have used this workspace in only one pipeline. But we do have different pipeline for different workspaces.

     

    The issue is not only for one visual, it's happening for various visuals except the places where Python script visuals is used.

    visual breaking here are normal TABLE visual

    • Brunner_BI's avatar
      Brunner_BI
      Icon for Impactful Individual rankImpactful Individual
      2 years ago

      So if you download the report in UAT you can confirm it is connected to the dataset you expect it to be connected to?

       

      Generally this seems to be hitting the limitations of DirectQuery by what you are doing inside the report, maybe month ending is a calculated column or so.

      Make sure to read through the limitations and then analyze what could be the cause. It could be your data model or a calculated column or something with Auto Date/Time?

       

      https://learn.microsoft.com/en-us/power-bi/connect-data/desktop-use-directquery#other-directquery-limitations

      • PRAVISH_SARAVAN's avatar
        PRAVISH_SARAVAN
        Frequent Visitor
        2 years ago

        sure, will go through the shared link on DirectQuery Limitations.
        Also, the MonthEnding is not a calculcated column, it's pulled form the created Dataflow.

        The strange reason was, until last friday(3/11/2023), the reports and deployments were wokring fine, starting this monday(6/11/2023) the reports started to not load the visual in TEST workspace. 

        Was there any new release happened over last weekend? Could that be possible route cause, if so any release udpate we can expect this weekend?