Get certified for free when you join Fabric Data Days 2026 and dive into Fabric, Power BI, SQL, AI, and other essential data skills.
Join nowData Days is here! Join us now for 60+ days of learning, challenges, and connection. Learn more
Hello guys,
I'm creating a semantic model that will be used as a source for other report. So here is the step taken:
1. Create report from Power BI Desktop, connect to Fabric Lakehouse, and use Import Mode.
2. In this "report", I only playing the query and relationship across tables, and create a "common" DAX measures.
3. Publish this to Fabric workspace, lets say MM_Datawarehouse. As expected, the workspace will contained Report and Semantic Model: lets say SM_Logistic (Report and Semantic Model)
4. I've created 2nd report (also with PBI desktop), lets say the name is "ReportLogistic" and this time Get Data from Power BI Semantic model, which choose the Semantic Model that just being uploaded (SM_Logistic)
5. Here is the place when I create the report, visualizations, and measures for my report.
6. Published this report to Workspace, which expectedly, only 1 object in that workpace, which is report.
7. The semantic model then have scheduled refresh daily.
Question is, is my report "ReportLogistic" will then get the latest update data when it is open ?
If I take a look at Semantic model Refresh history, it is refreshed correctly. And if I check the backend, from SQL endpoint, query, the table is updated.
But it seems my Report is not taking the latest refreshed. It only displayed the data from the last published.
Even when I check the report which build the semantic model, SM_Logistic , if directly open this report in Fabric and look at some of the table which I can tell it is updated or not, it will dusplay the last update of the published.
May I know what happened ? is there any step that I missed ?
Is it beause the semantic model create from the 1st report, is by Import Mode? so the semantic model only contained the last published of the report ?
Means even when the Semantic Model created is refreshed daily, the data it self is not updated / pull new data from the source connection ?
If I open again the 1st report, from Power BI destop, and refreshed. which if I check the data then up to date (before refresh is not up to date). I believe if I publish this again to the Workspace, all report is correct. But this is not in my mind or my plan.
I do not want to open my 1st report then refresh everytime and publish everytime, because this report is only to design and create the model. I will only open this 1st report, if I want to make changes of the model, or create new common measure so all report connected get easily use it.
What I had in mind is, after the semantic model created and published, I will just schedule this to refresh daily. then all report connected to this semantic model (Live connection) should get the latest data from the source whatever underlying data in that model connected with.
Thanks,
Solved! Go to Solution.
Your setup might be correct conceptually but because SM_Logistic(in your case) uses Import mode, data only updates when the model itself refreshes successfully in the service & not when reports connected to it are opened.
The issue is that your 1st report’s semantic model (SM_Logistic) is Import mode, so it holds a cached copy of data.
When you schedule a refresh in PBI Service, it only refreshes that semantic model’s internal cache, it does not automatically requery your fabric lakehouse unless the dataset refresh actually runs successfully from service (and uses a valid gateway or fabric conn)
In your case:
The 2nd report (ReportLogistic) is live connected to SM_Logistic
Therefore it only shows whatever data is in SM_Logistic’s cache
If SM_Logistic’s refresh is not actually pulling new data from the lakehouse (due to Import mode or connection issue), your live report will not show updated data
To fix this:
Ensure the SM_Logistic dataset’s scheduled refresh in the Power BI Service actually connects and pulls data from the Lakehouse (check refresh logs)
Or, switch SM_Logistic’s connection to direct lake/directquery mode so it always queries the latest Lakehouse data
Hi @VoltesDev,
As @Vinodh247 mentioned, if your first model is in import mode, it won't automatically refresh with data from the lakehouse. For that, you need to use DIrectLake mode.
TO make a DirectLake mode model, instead of using the SQL endpoint, use the OneLake catalog button in PowerBI Desktop and select the lakehouse tables from there.
If you found this helpful, consider giving some Kudos. If I answered your question or solved your problem, mark this post as the solution.
Proud to be a Super User! | |
Hi @VoltesDev ,
May I ask if you have resolved this issue? Please let us know if you have any further issues, we are happy to help.
Thank you.
Hi @VoltesDev ,
Thank you for reaching out to Microsoft Fabric Community.
Thank you @Vinodh247 @tayloramy for the prompt response.
I wanted to check if you had the opportunity to review the information provided and resolve the issue..?Please let us know if you need any further assistance.We are happy to help.
Thank you.
In fabric workspaces, datasets connected to a Lakehouse (especially in Import mode) use a builtin Fabric connection. You will not see a gateway setting because Fabric handles that internally. However, your dataset still needs explicit credentials to connect back to the Lakehouse. If the connection was created locally in PBI desktop and then published, those credentials do not always persist in the service.
Go to the SM_Logistic semantic model
Open Settings -> data source credentials.
If it says credentials missing or invalid, re-enter them and test the connection.
Then trigger a manual refresh.
Once that works, your scheduled refresh should also pull fresh data from the Lakehouse.
That usually fixes the “data not updating” symptom.
Hi @VoltesDev,
As @Vinodh247 mentioned, if your first model is in import mode, it won't automatically refresh with data from the lakehouse. For that, you need to use DIrectLake mode.
TO make a DirectLake mode model, instead of using the SQL endpoint, use the OneLake catalog button in PowerBI Desktop and select the lakehouse tables from there.
If you found this helpful, consider giving some Kudos. If I answered your question or solved your problem, mark this post as the solution.
Proud to be a Super User! | |
Your setup might be correct conceptually but because SM_Logistic(in your case) uses Import mode, data only updates when the model itself refreshes successfully in the service & not when reports connected to it are opened.
The issue is that your 1st report’s semantic model (SM_Logistic) is Import mode, so it holds a cached copy of data.
When you schedule a refresh in PBI Service, it only refreshes that semantic model’s internal cache, it does not automatically requery your fabric lakehouse unless the dataset refresh actually runs successfully from service (and uses a valid gateway or fabric conn)
In your case:
The 2nd report (ReportLogistic) is live connected to SM_Logistic
Therefore it only shows whatever data is in SM_Logistic’s cache
If SM_Logistic’s refresh is not actually pulling new data from the lakehouse (due to Import mode or connection issue), your live report will not show updated data
To fix this:
Ensure the SM_Logistic dataset’s scheduled refresh in the Power BI Service actually connects and pulls data from the Lakehouse (check refresh logs)
Or, switch SM_Logistic’s connection to direct lake/directquery mode so it always queries the latest Lakehouse data
Hi Vinodh,
May I know how to change my SM_logistic dataset (semantic model) connects and pulls data from the Lakehouse then ?
Yeah I guess, this is the issue. I initialy thought by published this to Fabric and schedule it, it will requery and pull again the fresh data.
If I take a look at the Semantic Mode -> Settings, it doesn't have Gateway and cloud connections ?
To switch to direct lake/ direct query, I'm just afraid that I need to rebuilt all the measures, so if possible this is my second options.
Kindly advice.
Thanks,
Hi @VoltesDev ,
Since your semantic model
(SM_Logistic) was built using Import Mode, it only contains a snapshot of the data at the time of publish. Even if you've scheduled refreshes, the model won't pull new data unless the dataset itself is configured to re-import from the Lakehouse.
To fix this:
Make sure the semantic model has valid credentials and a working scheduled refresh that connects back to the Lakehouse.
Confirm that your report (ReportLogistic) is bound to the refreshed semantic model-you can check this in the lineage view.
If needed, you can rebind the report to the correct dataset using Power BI Studio in VS Code.
Once the semantic model is refreshing properly, all connected reports should reflect the latest data automatically.
Thanks,
Sai Teja