Forum Discussion
Data update in report combining Direct Query and OneLake Data Hub (Power BI Semantic Model)
Hi,
I am facing a difficulty with updating a report, which uses onelake datahub and direct query.
I created a report R2, which uses OneLake Data Hub(Power BI semantic model) S1 connector and added Direct Query connected data S2 from another source, which forced me to create "local model".
I thought that if I now did some semantic model updates on the S1 (and published it), it would show up on the R2 report online immediately. But it seems that it doesn't.
When I e.g. change the data type of a certain column in the S1 and I want this change to be visible in the R2, I have to open R2 in the desktop, refresh it and then publish it again.
This seems stupid. I would have expected, that a change to S1 after publishing would immediately take effect on the R2 online (same as if I had just a regular thin report only using the S1). But it doesn't.
Any advice? I don't want to spend time on opening the "thin reports" that combine multiple sources and refreshing them and republishing every time, I have some change done in the main report/semantic model.
Any advice? or is this just the stupid native behavior?
I would have expected, that a change to S1 after publishing would immediately take effect on the R2 onlineThat is only true for data, and for varying interpretations of "immediately".
Meta data changes need to be propagated.
8 Replies
This behavior is indeed a limitation when using composite models that combine DirectQuery and a live connection (like OneLake Data Hub's semantic model). Here’s what’s happening and some options to consider:
1. Local Model Constraints: When you add DirectQuery sources to a report connected to a semantic model, Power BI creates a local model. Unfortunately, the local model does not automatically sync with updates made to the original semantic model. This limitation requires reopening, refreshing, and republishing the report to reflect any updates made to S1.
2. Alternative Approaches:
- Use DirectQuery Only: If feasible, consider sourcing all data via DirectQuery to avoid needing a local model. This allows your report to respond to live changes in connected data sources.
- Automate Desktop Refresh and Republish: Power Automate Desktop or other scheduling tools can help automate opening, refreshing, and publishing reports. While it doesn’t replace live syncing, it can minimize manual intervention.
- Dataflows as Intermediate Layer: If it’s feasible, use Dataflows for common data sources in S1 and S2. Then, have R2 pull from these Dataflows, which allows faster refreshes and somewhat centralizes updates.Unfortunately, native behavior currently requires manual intervention for reports with mixed live and DirectQuery sources. Hopefully, future updates will improve integration across composite models with these different data sources.
I would have expected, that a change to S1 after publishing would immediately take effect on the R2 onlineThat is only true for data, and for varying interpretations of "immediately".
Meta data changes need to be propagated.
