Skip to main content
cancel
Showing results for 
Search instead for 
Did you mean: 

Score big with last-minute savings on the final tickets to FabCon Vienna. Secure your discount

Changes to DAX formula not picked up in deployment pipeline.

We have a really simple DAX measure which is used to dyanmically set the title on a page in a report:

 

Page Title = MIN(Indicators[indicator_code]) & "  " & MIN(Indicators[indicator_description]) & UNICHAR(10) & "Trends"
 
We the client viewed the report, they noticed that there was an additional space between indicator code and indicator description.  So we remove the rogue space in measure in the version of the report / semantic model in the DEV workspace:
 
Page Title = MIN(Indicators[indicator_code]) & " " & MIN(Indicators[indicator_description]) & UNICHAR(10) & "Trends"
 
However the change to the DAX measure was not flagged by the deployment pipeline so we didn't get a visual cue that there were differences between DEV and TEST.  Seems there is a "bug" in the logic which compares semantic models across workspaces that means changes to white space is not being picked up?
 
Note - we chose to deploy the semantic model from DEV to TEST and then from TEST to PROD and the change was propogated, but it was strange that it was never flagged in the UX.
 
In this case, this was a simple cosmetic change.  But my fear is that this could result in something more fundamental not being flagged in the UX and therefore changes not being picked up and propagated.
Status: Investigating
Comments
Anonymous
Not applicable

Hi,@Barry_Smart .I am glad to help you.
I understand your concern.
Here are my inferences:
A change in measure in a semantic model is not usually considered an architectural change. Architectural changes usually involve changes to table structures, columns, relationships, etc., whereas changes to measure are more of a logical adjustment. And in your case, it doesn't involve the name of the measure being changed.

And according to the documentation provided by Microsoft for pipeline deployment, the comparison involves more of an architectural change to the mod (this is only my understanding of the documentation)
Compare content in different stages before deployment - Microsoft Fabric | Microsoft Learn

I would highly recommend creating a support ticket to get more help.

The Link of Power BI Support:

Microsoft Fabric Support and Status | Microsoft Fabric


 

vjtianmsft_0-1737688877559.png

 

I hope my suggestions give you good ideas, if you have any more questions, please clarify in a follow-up reply.
Best Regards,
Carson Jian

gabrielPerson
Frequent Visitor
Hi Barry (couldn't tag you, sorry), I encountered a similar issue today with deployment pipelines not detecting certain changes in calculated columns. In my case, the change was as minor as removing a single blank space in an IF statement inside a COALESCE function. After running some tests, I noticed this behavior only occurs in specific scenarios like yours and mine. However, I found a workaround: if I add or modify a comment in the measure or calculated table (even a small change to an existing comment), the pipeline recognizes the update and successfully deploys it to production. I confirmed this using Tabular Editor. While this isn’t an ideal solution, it serves as a temporary workaround for this odd behavior in deployment pipelines.