Forum Discussion
Fabric - semantic model - refresh
I found the issue
I had one view in the warehouse with two columns having identical names where one of them had a capital letter
I tried to refresh the warehouse from fabric gui and received the duplicate column name error message
It is fully possible (at least from sql endpoint) to create a view with two columns names aa and Aa for example. However its not supported in semantic models and semabtic model will not give you an indication of this issue
Hi
this is urgent and frustrating
shall I create a new forum topic for this issue specifically?
I think I will recommend to create a support ticket https://support.fabric.microsoft.com
Also, I think there are two separate concepts here:
- Edit the schema of the model ("Edit Tables"). https://learn.microsoft.com/en-us/fabric/get-started/direct-lake-edit-tables
- Refresh the data content. https://learn.microsoft.com/en-us/fabric/get-started/direct-lake-overview#refresh
I believe the first one is the cause of your issues?
The latter can be done automatically, in that case there would be no need for your users to refresh.
Perhaps you will find your issue listed as a Known issue here: https://learn.microsoft.com/en-us/fabric/get-started/fabric-known-issues#currently-active-known-issues
Similar topics:
Using Tabular Editor to edit semantic models may be an alternative.
Or using Power BI desktop (import mode).
Are you using Lakehouse or Warehouse?
Anyway, using views is bad for performance in direct lake semantic models. Views make the report queries fall back to DirectQuery (which has lower performance than Direct Lake). So the recommendation is to use tables, not views, if you wish to use Direct Lake mode.
I found the issue
I had one view in the warehouse with two columns having identical names where one of them had a capital letter
I tried to refresh the warehouse from fabric gui and received the duplicate column name error message
It is fully possible (at least from sql endpoint) to create a view with two columns names aa and Aa for example. However its not supported in semantic models and semabtic model will not give you an indication of this issue
That's interesting! Thanks for sharing. So parts of Fabric is case sensitive while other parts of Fabric is not case sensitive. That's good to know. I think there should be a warning indicator or error message for these kind of issues.
Thanks for getting back
My biggest issue rigth now is the first one ("Edit tables")
I use Warehouse as a source but there is a mix of tables (warehouse) and views (sourced from delta tables in lakehouse). I am aware about the direct lake functionality and I will soon only have tables in Warehouse as source. However, the semantic edit table refresh shoudl still work (pick up the updated model from warehouse).
I will create a support ticket.Can you please guide me on how I do that or send me a more specific link? What permission do I need?
I don't have experience with it, but I think scroll down to Power BI -> Get customer support https://support.fabric.microsoft.com
I think you should also check with IT / Microsoft 365 Admin in your organization to see if you have some special considerations regarding support tickets, how much data you can share with the MS support personnel, etc.
I think the issue you are encountering is a known issue and current workarounds seem to be to use Tabular Editor or Power BI Desktop (import mode).
thanks
Import mode works but I need to create a general model that van be used for mutiplöe reports and be consumed directly from excel. Can you acheive this with Power BI Desktop?
regarding tabular editor - can ypou publish a semantic model or a general model from that tool?