Help with Semantic Model connection

[GlibAvrutin]

Hello guys, I'm new to the PowerBi service so please don't judge too harshly  🙂

The thing is that I've updated some kql RLS stuff in my dev fabric workspace and now trying to refresh the semantic model in the same workspace, but it fails because it points to the different kql database (not in my current dev workspace, error message attached). I've tried changing cloud connection but it didn't work, asked AI and searched some topics on this forum and couldn't find any good pieces of info. 

Data source error: Bad request: Semantic error: Error in row_level_security query for database("fba6f93a-
102c-4b5d-88d5-5043df92810d").table("_MV_DimResponse"): the following operators were blocked:
'summarize' (only the following operators are allowed: 'as', 'distinct', 'extend', 'join', 'limit', 'project', 'project-
away', 'project-keep', 'project-rename', 'project-reorder', 'union', 'where'). Error = Semantic error: Error in
row_level_security query for database("fba6f93a-102c-4b5d-88d5-
5043df92810d").table("_MV_DimResponse"): the following operators were blocked: 'summarize' (only the
following operators are allowed: 'as', 'distinct', 'extend', 'join', 'limit', 'project', 'project-away', 'project-keep',
'project-rename', 'project-reorder', 'union', 'where'). Code = General_BadRequest. Type =
Kusto.Data.Exceptions.SemanticException. x-ms-activity-id = e820ca64-a257-4fdc-92cb-e18fc1da7633. x-ms-
client-request-id = KPBI;cbbdb9d9-a5e3-473c-886e-8b1cb882c188;cffb9d11-fd6b-4657-8a85-
0220dda9b1f2;c3cdb7fe-cf52-4977-bdc8-0f8c62388694 .. The exception was raised by the IDbCommand
interface.

Thank you very much in advance for any help! 

[v-sgandrathi]

Hi @GlibAvrutin,

The issue occurs because the semantic model is still set to connect to the PROD KQL database, and simply changing the cloud connection does not update the database reference in the model. In Microsoft Fabric, the KQL database details are specified in the data source connection within the semantic model.
To resolve this, go to your DEV workspace, open the semantic model, select Edit, and then open Transform data (Power Query). Check the KQL connection step for the cluster URL and database name, which likely point to the PROD database. Update these to use the DEV KQL database, save, and republish the model. After publishing, go to Settings -- Data source credentials to re-authenticate if needed and refresh the model.

If the model cannot be edited (such as when deployed via pipeline), you will need to redeploy it with environment-specific rules or create a new semantic model in DEV connected to the DEV KQL database. The main point is that the database reference inside the model must be updated; changing only the cloud connection will not redirect it to DEV.

Thank you.

[GlibAvrutin]

Hi @Mauro89, Thank you for your answer. Yes there were issues with KQL query in all of my fabric workspaces (dev, staging and prod) and I managed to solve them for dev database, but the semantic model located in dev workspace, which was created and configured not by me, points to the db in prod workspace, not in dev, as the id in the error message suggests. 

[Mauro89]

If you just solved the issues in DEV workspace/database, then maybe the KQL code differs between DEV and PROD. And as the semantic model takes the PROD KQL DB maybe this could be worth checking and comparing. 
I also found this interesting post which might help if there are issues with related tables https://stackoverflow.com/questions/67092083/how-to-recover-from-a-broken-row-level-security-policy#...'

Best regards!

[GlibAvrutin]

Yep, that's right, and basically the question is how to make semantic model point to the dev database instead of prod. I tried creating new cloud connection in the model parameters tab,  but it didn't give any results (screen attached) 

Thank you for support

[Mauro89]

Hi @GlibAvrutin,

welcome to Power BI and the community 🙂 and no worries, the community is here to help, not to judge  😊

So your error message indicates some trubble with the KQL query which is underlying in your KQL database. Seems like you have some semantic issues as the error states the "summarize" operator is not stated correct.

Therefore it would be a good first point to start investigate the KQL query. If you can, share the query with the community then we can dive deeper. 

Best regards!

PS: If you find this post helpful consider leaving kudos or mark it as solution