Power BI refresh job task

[alesventus]

I have tried Databricks job task to refresh power bi dataset and I have found 2 issues.

1. I set up tables in Power BI Desktop using Import mode. After deploying the model to Power BI Service, I was able to download it as an Import mode model. However, after running the first Databricks job with a Power BI task, I could no longer download the report in Import mode—only as a copy with a direct connection. I suspect this happened because the Databricks job modified the Power BI model via XMLA.

Databricks job setup

2. I was unable to refresh a Power BI model that included a Rename Column step in Power Query. When I removed this step, the refresh completed successfully.

 Error message:

Column month_number was renamed in the Power BI Desktop (Transform data - Steps) to Month Number.

Does anybody have the same experiences? Is there any workaround for this?

Thanks

[emma_s]

Hi, 

1. In your job did you update the power bi query mode to import mode before you ran it?

2. If changing to import mode doesn't fix this issue, have you tried renaming in the semantic model rather than in the transform data steps?

[alesventus]

I set this option to Import before the first job run. I tested the scenario again and observed the same result: before the job runs, the option to download in Import mode is enabled; after the run, it becomes disabled.

Regarding renamed columns in Power BI Power Query, I couldn’t find any mention of such a limitation in the documentation. It might be an undocumented feature/bug.

[emma_s]

Can you send a screenshot of the refresh power BI task in the jobs UI within Databricks please?

[alesventus]

Databricks job contains just one task.

There is only one source of the data - view based on delta lake table in Unity Catalog.

Cluster is serverless sql warehouse.

This is the result of the job: