Forum Discussion
Problem with semantic model refresh after data transformation in Data Flow Gen2
Hello DALI007 ,
check out this post, maybe it'll help :
Re: How can I update the PowerBi semantic model sc... - Microsoft Fabric Community
Hello,
Exactly I used a Dataflow Gen2, a semantic model and a Lakehouse.
To my knowledge, with the default settings, a Direct Lake dataset is automatically refreshed when the data is loaded into a Lakehouse.
Yes, Initially I published manually the dataFlow Gen2 which made me the creation of the table in the lakeHouse.
In a second time I made a transformation on the table at DataFlow Gen2 (Adding a column "type_site" in the table).
my problem how to replace the existing table automatically or manually if the shemas of the table latter changes (in my case : addition of this new column for example) ? Thank you
Is the issue that the new column is reflected in the Lakehouse but not in the automatic semantic model? Or that the new column is not reflected in the Lakehouse at all?
Yes, the new column is not reflected in the Lakehouse at all
A Dataflow Gen2 doesn't change the schema of a table in the destination at the moment.
You can either create a new table by going through the "data destination" configuration for your query and defining a new table, or make the changes in your backend (your LakeHouse) to alter the schema of a table and then update your dataflow gen2 to load the data to the updated schema. In both scenarios you'll have to go into the dataflow gen2 and update the configuration for the data destination.
hope this helps!
Hi miguel ,
I have a similar problem.
I initially created a schema in a warehouse. I use this warehouse to create a report using PowerBi Desktop. From this point on, I can no longer see the schema changes I make in the warehouse. If I add a column to a dw table, I can't see it in the PowerBI model.
Is this normal?
How can I update the PowerBi semantic model schema when I make changes to the source structure?
What's more, even when I delete the table from the semantic model, and add it again, I get back exactly the same structure as before, without the new column...