Pushed semantic model is not listed in dataset list while using semantic-link-sempy Python package

[Karel]

Hi all,
We set the dta transfer from an external data source which creates pushed semantic model via api connection in My workspace. Pushed data are very limited to be used for custom reports and it's not possible to edit pushed tables in any other way than adding new Measures. We need to adapt pushed tables and add some calculated columns etc.
So we tried to create a new Lakehouse and copy all the data from pushed semantic model to the new one with usage of semantic-link-sempy Python package. The current problem is that pushed semantic model is not listed as dataset while using fabric.list_datasets() so fabric.read_table() can't be used to read and copy data to a new model.
Are there any specific limitations for pushed semantic models? Or is there any other way hot to manipulate data in this situation?
Thank you in advance for any hint.  

[Karel]

@Anonymous 
Hello,
Thank you for your answer. I've seen the notice about some Pushed dataset limitation but I was hoping this is not so strict. We don't need real-time data streaming but our goal is to have data pushed directly to Power BI where all the data transformation can be done. The refresh process is maintained on the side which connecting to the Power BI. It seems Power BI can't handle this in a simple way so we need to search for other tool in this case.

[lbendlin]

our goal is to have data pushed directly to Power BI where all the data transformation can be done.

 those things are mutually exclusive.

[Anonymous]

Hi @Karel ,

Firstly, it's important to understand that pushed semantic models in Power BI have specific characteristics and limitations. As you've observed, these models are designed primarily for near real-time streaming scenarios and are limited in their functionality compared to other types of datasets. Specifically, pushed semantic models are not intended for extensive custom report creation or editing beyond adding new measures. This is due to the nature of how data is stored and managed within these models. For more detailed information on the limitations and intended use of pushed semantic models, you can refer to the Push Datasets Limitations documentation.

Regarding the issue with fabric.list_datasets() not listing your pushed semantic model, this behavior aligns with the limitations of pushed semantic models. These models are managed differently within the Power BI service, which can affect how they are accessed and manipulated through external tools and APIs.

To work around this limitation and achieve your goal of adapting pushed tables and adding calculated columns, you might consider the following approach:

1. Export Data Programmatically: Instead of directly accessing the pushed semantic model, you could export the data programmatically from Power BI to a storage solution like Azure Blob Storage or Azure Data Lake.
2. Create a New Semantic Model: Use the exported data to create a new semantic model in Power BI Desktop or the Power BI service, where you have more flexibility to add calculated columns and perform other manipulations.
3. Automate Data Refresh: Set up an automated process to periodically refresh the data in your new semantic model, simulating a near real-time data flow.
   This approach allows you to bypass the limitations of pushed semantic models while still leveraging the real-time data they provide.

Best regards,
Community Support Team_Binbin Yu
If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.