Forum Discussion
Helper IIDeployment pipeline results in "Different from source"
Hi everyone,
I just set up deployment pipeline today. I only had production workspace before. I tried to deploy backwards from prod -> test -> dev to populate the pipeline with the most recent files. This is when the pipeline starting to flag differences (see below).
There is no issue if published from desktop directly, however as soon as I started deploying, whether from dev to test or prod to test etc. it would always result in differences (always the same missing line items in the screenshot below).
I have not tried delete everything from all workspaces and republish.
Hello TRADER083 ,
Yes, sometimes you can have these strange little things that happen, and the reasons for them can be multiple :- Model metadata: Power BI Desktop may generate slightly different definitions depending on the version or how the model was created.
- Calculated columns vs. imported columns: Calculated columns can have attributes like dataType or isNullable, which aren't always synchronized between workspaces.
- Pipeline doesn't synchronize everything: Deployment via pipeline compares the model's JSON files (schema), and even minor differences (like order or implicit attributes) are reported.
To be clean (because pushing from Prod to Dev isn't a best practice), I recommend:
- Reset the workspaces:
- Delete the datasets in Dev and Test.
- Republish from Desktop to Dev.
- Then use the pipeline to propagate to Test and Prod.
- This ensures that the schema is identical everywhere.
Don't hesitate to give a kudo if my answer helped you.
Have a nice day,
Vivien
6 Replies
- vivien57
Super User
Hello TRADER083 ,
Yes, sometimes you can have these strange little things that happen, and the reasons for them can be multiple :- Model metadata: Power BI Desktop may generate slightly different definitions depending on the version or how the model was created.
- Calculated columns vs. imported columns: Calculated columns can have attributes like dataType or isNullable, which aren't always synchronized between workspaces.
- Pipeline doesn't synchronize everything: Deployment via pipeline compares the model's JSON files (schema), and even minor differences (like order or implicit attributes) are reported.
To be clean (because pushing from Prod to Dev isn't a best practice), I recommend:
- Reset the workspaces:
- Delete the datasets in Dev and Test.
- Republish from Desktop to Dev.
- Then use the pipeline to propagate to Test and Prod.
- This ensures that the schema is identical everywhere.
Don't hesitate to give a kudo if my answer helped you.
Have a nice day,
Vivien
TRADER083Thanks Vivien, unfortunately clearing the workspace and doing a clean slate upload to Dev -> Test did not work. Perhaps the indexing and sorting I was doing in PQ is not properly supported.
- vivien57
That's strange.
To summarize, you only had one production workspace.
- You synchronized it with Git and pushed the developments to it.
- You then created a new workspace, linked it to Git, and pulled what you just pushed from the production workspace.
- Then you created a deployment pipeline between the two workspaces.
Have a nice day,
Vivien
- TRADER083
The issue was not resolved by cleaning all workspaces and doing a refresh load to Dev -> Test -> Prod.
However, I went back to the file and located calculated column with the issue and updated the data type from whole numbers to decimals and this resolved the issue.
v-saisrao-msftCommunity Support
HI TRADER083,
Thank you vivien57, for your insights.
Use Fabric deployment pipelines to move content from development to test to production, deploying either all items or only selected ones. Make sure deployments flow in a single direction, and review all changes in the change review window.
Power BI implementation planning: Deploy content - Power BI | Microsoft Learn
Thank you.