Forum Discussion
Semantic model does not load all data from Dataflow Gen 1
Hi, community.
We are facing a behaviour/error that we can not explain regarding semantic model, in import mode, connected to a Dataflow Gen 1 in a workspace running on a Fabric Capacity.
The scenario we are experiance:
- Checking the Dataflow, all data we need is there. The data is far less that 1GB. (aprox 500MB)
- Connecting Power BI Desktop to the Dataflow, randow rows are missing from the dataset that we know exisits in the Dataflow. Some of them from 2025.
- Limiting the data that exists in the Dataflow (change filter on year to ingest from 2020 to 2024) and refresh the Dataflow. Now all data from 2024 and up until today are loaded into the Semantic model. Dataset now less that 300MB.
Are there anything in how semantic models load data from a Dataflow Gen 1 that could explain this behaviour when the dataset is clearly inside the limitations of the Semantic model, and the Dataflow for that matter?
Do you need anymore info or have question, please just ask!
Regards
Jon Terje Tjønn
If you have a Pro license you can open a Pro ticket at https://admin.powerplatform.microsoft.com/newsupportticket/powerbi
Otherwise you can raise an issue at https://community.fabric.microsoft.com/t5/Issues/idb-p/Issues .If it is a consistent number of rows that makes me think that the issue is with the load and the rows are being errored out for some reason. That should be very evident in Transform data though. Any inconsistent number of rows would lead me in the direction of resource constraints.
IMO you are firmly in the open a ticket with Microsoft support arena.
Please mark this post as solution if it helps you. Appreciate Kudos.
10 Replies
Hello JonTerje
Thank you for reaching out to Microsoft Community Forum.
The issue where a Semantic Model (Import Mode) does not load all data from a Dataflow Gen 1 in Fabric Capacity could be caused by several factors, including row limitations, refresh issues, and potential caching.Verify whether the Dataflow refresh has completed successfully, ensuring that all data is available in the Dataflow and properly updated in the Semantic Model. Dataflow Gen 1 in Import Mode updates data only when the Semantic Model is refreshed. If data is missing, manually refresh the dataset in Power BI Service to check if the rows appear.
If specific rows, especially from 2025, are missing, review the Dataflow preview in Power BI Desktop to confirm all rows are visible. If data appears in the dataflow but is missing in the model, it indicates potential issues with the Semantic Model import process.
As mentioned testing with data limited to 2020-2024 loaded successfully, it is due to a refresh limit, model constraint, or Power BI bug related to 2025 data. Try loading only 2025 data into the Dataflow, and if it fails to load, consider issues with date filtering or a possible bug, whereas successful isolation may suggest challenges in processing larger datasets with mixed year ranges
If my response has resolved your query, please mark it as the Accepted Solution to assist others. Additionally, a 'Kudos' would be appreciated if you found my response helpful.
Thank you!
Hi and thank you for your answer, v-karpurapud.
The workspace where we see this issue is running on a Fabric F64 Capacity with enough available resources.
Scheduled refresh is set up for both Dataflow Gen1 and the Semantic Model, with plenty of time between the refresh updates. Both refresh runs successfully. However the data is incorrect with several rows just missing in the semantic model, event though the Dataflow Gen1 contains the data once I go into it and check.
What are the row limitations for both these objects?
After manually limiting the data to 2024 -> in the Dataflow and refresh both Dataflow and Semantic Model manually, the data is correct.
Next scheduled refresh, the data was again wrong.
I have never tried to load just historic data (2020 - 2024).
Opening a ticket with Microsoft might be the best next step here.
Regards
Jon Terje
Scheduled refresh is set up for both Dataflow Gen1 and the Semantic Model, with plenty of time between the refresh updates.Consider using Power Automate - it has triggers for "When a dataflow refresh completes" and actions for "Refresh Semantic Model"
If you have a Pro license you can open a Pro ticket at https://admin.powerplatform.microsoft.com/newsupportticket/powerbi
Otherwise you can raise an issue at https://community.fabric.microsoft.com/t5/Issues/idb-p/Issues .This is the first time I have heard of this.
In your model, in Transform data did you check the column statistics (profile) to see if there were any errors?
How are you validating that the rows are missing in your table?
Please mark this post as solution if it helps you. Appreciate Kudos.
I filter for spesific data rows and see that the same filter in the Dataflow (Spesific olumn | Begins with | Value) gives a different result that the same filter in Power BI Desktop (Drag column to canvas | advanced filtering | starts with | Value)
Regards
Jon Terje Tjønn
There is a row limit on the canvas visuals, although that shouldn’t really matter when you are specifically filtering it down.
However, I would step back and try and narrow down the problem and see if it is a consistent problem. Start with confirming what the row count should be and then check that against the row count on the table in your model to see if it is actually different. If it is then I would try to recreate the problem to see if I get to the same missing number of rows each time.
Next, I would go to Transform Data and check some column profiles and look for errors.
Please mark this post as solution if it helps you. Appreciate Kudos.
