Multiple Semantic Model created by Lakehouse - Cannot delete

[Timbaaaal]

Hi,

I published a dataflow Gen 2 yesterday which take data from 4 tables in one Lakehouse in order to create a new table in another LH (lets call it Gold).

The gold LH was new and never had any table. I think I might have created it while doing the destination mapping in the dataflow.

I now end up with 4 semantic model auto created (with no table) + the gold LH which had the table but no semantic model (Creation of SQL analytics endpoint/Semantic model was refreshing and my guess, stuck).

I deleted the gold LH, but cannot delete the 4 semantic model... Any thoughts?

Workspace is F2 capacity/West-Europe and is part of a development pipeline, but those items never got deployed.

PS: I am able to move them in a folder

[v-csrikanth · Accepted solution]

Hi @Timbaaaal 
Could you please navigate to the workspace settings and you can delete orphaned semantic models. 
Your issue will get resolved.

If the above information is helpful, please give us Kudos and mark the response as Accepted as solution.
Best Regards,
Community Support Team _ C Srikanth.

[Timbaaaal]

This works like a charm. Thanks!

[v-csrikanth]

Hi @Timbaaaal 
It's been a while since I heard back from you and I wanted to follow up. Have you had a chance to try the solutions that have been offered?
If the issue has been resolved, can you mark the post as resolved? If you're still experiencing challenges, please feel free to let us know and we'll be happy to continue to help!
Looking forward to your reply!

If the above information is helpful, please give us Kudos and mark the response as Accepted as solution.
Best Regards,
Community Support Team _ C Srikanth.

[v-csrikanth]

Hi @Timbaaaal 
Sorry for the late response and inconvenience caused.
The best way I can suggest to follow the steps to resolve your issue. 

1. Delete the Associated Dataflow:

   • The creation of multiple semantic models is likely linked to the Dataflow Gen2 that was set up to map data to the target Lakehouse. Deleting this dataflow can remove the unintended semantic models.

2. Verify Dependencies Before Deletion:

   • Ensure that the dataflow isn't being utilized elsewhere to prevent unintentional data loss.

3. Refresh the Workspace:

   • After deletion, refresh the workspace to confirm that the semantic models have been removed.

If you are still facing any challenges, kindly provide further details, and we will be happy to assist you.

If the above information is helpful, please give us Kudos and mark the response as Accepted as solution.

Best Regards,
Community Support Team _ C Srikanth

[Timbaaaal]

Hi @v-csrikanth 

Thanks for getting back.

I deleted the dataflow on 2-17-2025 and as of today, the semantics model are still there.

Please see my last comment, I mentionned what I did prior the deletion (of the dataflow) which probably is the reason why they didnt got deleted.

[v-csrikanth]

Hi @Timbaaaal 
Thanks for using the Microsoft Fabric Community.
I replicated the scenario locally and was able to resolve it. The solution is deleting the dataflow that was originally created to map the destination of the Gold Lakehouse. Once the dataflow was removed, the unnecessary semantic models were no longer present, and the process functioned as expected. This suggests that the issue was caused by the initial dataflow's destination mapping, which, when deleted, resolved the problem.

If you found the above information helpful, we kindly request you to give us a Kudos and mark the response as the Accept as Solution.

Thank you,
Cheri Srikanth

[Timbaaaal]

Hi and thanks for looking into this.

I just deleted the dataflow, but the 4 semantic model (default) are still there.

This might due to the following (which was done prior deletion of the dataflow):

1. I was able to create a new Lakehouse with the exact same name

2. I reuse the same dataflow by updating the destination mapping to the new lakehouse

3. The old semantic model were moved to a different folder than the dataflow/new lakehouse

I have now also deleted the new lakehouse (which had the same name), but it still didn't delete those 4 semantics model.
However I am no longer able to create a new lakehouse with the same name. It still let me click on Create (Normally it doesn't), but I get error with "Lakehouse Id is not valid".

This is new, not sure if it because I deleted the dataflow or the last lakehouse..

Expected behaviour:

Current behaviour:

When clicking on Create:

Error Details:

[StrategicSavvy]

hi, you might be dealing with this isse:

https://community.fabric.microsoft.com/t5/Data-Engineering/HOW-TO-DELETE-default-semantic-model/m-p/...

it's said it is not possible to delete default semantic models

[Timbaaaal]

Hi, thanks for sharing.

I did came across this one as well. But in this case, the person is looking to delete the semantic model while keeping the lakehouse. looks like Microsoft doesn't want to have a Lakehouse without a semantic model (default), so this is expected behaviour.
In my case, I already deleted the original Lakehouse, but the SM didn't get deleted and on top of that I had multiple created when the Lakehouse got created by the dataflow.

But like people mentionned on the other post, underlying issue is that we should definitly have a way to delete artifacts in the workspace. But also full visibility (Staging, soft delete etc.) If they don't want us to mess up with dependencies, then they should put in place better control/visibility of dependencies, like they did with objects in Dynamics/Power Apps Solution.