Forum Discussion

ToddChitt's avatar
ToddChitt
Super User
1 year ago
Solved

Semantic Model Refresh pipeline task shows no tables

I have a simple Fabric Pipeline task that loads data into warehouse tables. The last step is a Semantic Model Refresh task. I have had this Semantic Model for months now and have refreshed it numerou...
data pipeline
  • Anonymous's avatar
    Anonymous
    1 year ago

    Hi ToddChitt,

    Thanks for reaching out to the Microsoft fabric community forum.

    This does sound like a regression in how the task fetches table metadata from the semantic model. We've seen a few similar reports come in, and it's currently under internal review. As you have already raised the issue in the Idea Forums, in the meantime, here are a few steps that have helped others work around it:

    * Try disconnecting and reconnecting the semantic model in the task click the pencil icon next to the Semantic model field and re-select the model manually. Then click Refresh again in the table section to force metadata reload.

    * Ensure the model is fully published and not in a pending or draft state in Power BI especially if it's recently been modified.

    * As a temporary workaround, if you're not relying on selective table refresh, you can disable the table selection and let the refresh task operate at the full model level (by skipping table selection altogether). This is functionally equivalent to a full manual refresh.

     

    If I misunderstand your needs or you still have problems on it, please feel free to let us know.  

    Best Regards,
    Hammad.

michelmetran's avatar
michelmetran
Frequent Visitor
7 months ago

I'm having the same problem. It's very frustrating to build an entire pipeline to organize the data and, in the end, to "deliver" the updated data, the task simply fails!

 

I tried all the solutions suggested here and none succeeded.

To work around this, I abandoned the "Semantic Model Refresh" task and created a notebook that performs the same task using Sempy.

 

I suggest watching the video https://www.youtube.com/watch?v=ucQcdPhUjdo which shows several additional possibilities of Sempy...

Below is the code I used to update my Semantic Model...

import sempy.fabric as fabric

# Nome do modelo semântico e do workspace
# Se o notebook já estiver no mesmo workspace, o parâmetro workspace é opcional
workspace_name = 'Dominio'
dataset_name = "Domínio (gold)"


try:
    print(f'Iniciando atualização do modelo: {dataset_name}...')

    # Executa o refresh
    fabric.refresh_dataset(dataset=dataset_name, workspace=workspace_name)

    print('Solicitação de atualização enviada com sucesso!')

except Exception as e:
    print(f'Erro ao atualizar: {e}')