Edit tables of semantic model does not add tables

Hello,

I don’t think is temporally, it has been some days that we are struggling with this issue.

I’m fully working within the Fabric web portal.  A have a Lakehouse containing all the data, the views are basically used to de-normalize some data and to add business translations. Both Lakehouse and Datawarehouse are in the same Workspace (with Trial Fabric capacity).

The list of available tables/views is well refreshed, I can select the view, then clicking Confirm.  It runs for some seconds and the windows is closed.  Looking at the model, the tables are not added.  If I try several times, sporadically one of the tables is added, but not all times.

Regarding the choice of a Datawarehouse, we chose that option to be fully compatible with all CRUD operations via T-SQL.

We are using an automation tool that generates T-SQL procedures as part of our ELT process, to Create, alter the tables.  It also generates T-SQL to Update, Insert and Delete the data.

In my example, I only have views in the DWH. I know that using views will use DirectQuery. If performance is not good, I understood we will need to materialize the views logic into tables.

Thanks,

Martha

Here is some more info regarding XMLA endpoint for Direct Lake Semantic Models:
Learn about Direct Lake in Power BI and Microsoft Fabric - Microsoft Fabric | Microsoft Learn

Announcing XMLA Write support for Direct Lake datasets | Microsoft Fabric Blog | Microsoft Fabric

Editing your Direct Lake Datasets from Tabular Editor? Yes please! (youtube.com)

The YouTube video shows where to find the XMLA endpoint connection string:

Are you using the default semantic model, or have you made a custom semantic model? I think creating a custom semantic model is preferred:
Power BI Default Semantic Model or Custom; A Guide for Using in Fabric Environment - RADACAD

I think this is also true (or even required) if you want to use XMLA endpoint.

According to this blog: Controlling Direct Lake Fallback Behavior (fabric.guru)

"After updating the model using Tabular Editor, you will not be able to use the web modeling experience. You cannot use this method for an auto-generated default semantic model as it doesn't have XMLA endpoint."

I see. Yes, when wanting to work with T-SQL, stored procedures, etc. for all CRUD operations, I think it makes sense to use the Warehouse. 

To me, the issue you are facing seems to be a bug with the Fabric user interface.
Hopefully, the XMLA endpoint can be used as a workaround for adding/removing tables from the semantic model.
However, the preferred solution is that the bug in the user interface should be fixed.