Forum Discussion
"Combining Visuals from Two Semantic Models with Shared Slicers in Power BI
Hi all
In a single Power BI workspace, I have two separate semantic models and two reports — each report is connected to its own semantic model.
Now, I want to create a third report that includes visuals from both existing reports. These visuals need to be controlled by several common slicers — for example, Date, Country, Business Unit, etc. — which exist in both semantic models.
I know dashboards might be one way to combine visuals, but I’m concerned about slicer functionality across visuals sourced from different datasets. Ideally, I’d like to apply a single set of slicers that can filter visuals from both semantic models.
Of course, I’d prefer not to rely on dashboards since I have no experience with them. I’m open to any alternative approach that allows shared slicers across visuals coming from two different semantic models.
Thanks
Hi Nazdac2024
There’s currently no straightforward way in Power BI to create a report that combines visuals from two different semantic models (datasets) in the same canvas, with native filtering and interaction between them.
Power BI reports are designed to connect to a single model at a time. While it’s technically possible to create a composite model that connects to multiple datasets using DirectQuery to Power BI datasets, this approach requires:
-
Rebuilding a new model that references both existing datasets.
-
Creating shared dimension tables (like Date, Country, Business Unit) that both sources can relate to.
-
Handling potential limitations with relationships, performance, and model complexity.
Dashboards in the Power BI Service allow pinning visuals from multiple reports/models, but they don’t support cross-visual interactivity or shared filtering – so it’s a limited workaround.
If you’re open to rebuilding a lightweight model to aggregate key data from both sources, that would give you more control and flexibility – even more so if you can standardize the shared dimensions.
For more information please refer to the linked discussion:
https://community.fabric.microsoft.com/t5/Desktop/Combining-multiple-pbix-data-sets/m-p/1773781If this post helps, then please consider Accepting it as the solution to help the other members find it more quickly
-
4 Replies
Hi Nazdac2024
There’s currently no straightforward way in Power BI to create a report that combines visuals from two different semantic models (datasets) in the same canvas, with native filtering and interaction between them.
Power BI reports are designed to connect to a single model at a time. While it’s technically possible to create a composite model that connects to multiple datasets using DirectQuery to Power BI datasets, this approach requires:
-
Rebuilding a new model that references both existing datasets.
-
Creating shared dimension tables (like Date, Country, Business Unit) that both sources can relate to.
-
Handling potential limitations with relationships, performance, and model complexity.
Dashboards in the Power BI Service allow pinning visuals from multiple reports/models, but they don’t support cross-visual interactivity or shared filtering – so it’s a limited workaround.
If you’re open to rebuilding a lightweight model to aggregate key data from both sources, that would give you more control and flexibility – even more so if you can standardize the shared dimensions.
For more information please refer to the linked discussion:
https://community.fabric.microsoft.com/t5/Desktop/Combining-multiple-pbix-data-sets/m-p/1773781If this post helps, then please consider Accepting it as the solution to help the other members find it more quickly
-
Hi Nazdac2024 ,
Thank you for reaching out to Microsoft Fabric Community Forum. We appreciate the helpful contributions from danextian, Ritaf1983 and lbendlin for providing valuable insights on this topic.
To achieve this, you can create a composite model by using DirectQuery to connect both models in a single report. This approach allows you to display data from both models while maintaining interactivity with shared slicers. You’ll need to create shared dimension tables (like Date, Country, and Business Unit) that both models can relate to, ensuring slicers filter visuals from both models simultaneously.
Additionally, it’s important to test the performance of the DirectQuery setup, as larger datasets may impact query speed, and you may need to optimize the queries for better performance. Lastly, if Row-Level Security (RLS) is implemented in either model, ensure that it works as expected across the combined model.
I hope this will resolve your issue, if you need any further assistance, feel free to reach out.
If this post helps, then please give us Kudos and consider Accept it as a solution to help the other members find it more quickly.
Thankyou.
One report can connect to more than one semantic model. If you have the same common filters in both models then either connect the models*, pick a "winning" side and hide the other, or create reference tables that feed your slicers and control both models.
* NOTE: Keep the join column cardinality below 50K
Hi Nazdac2024
You can connect to multiple semantic models using DirectQuery. To manage and filter fact tables from both models, consider creating shared dimension tables—such as Date, Country, and Business Unit. If these dimension tables already exist in one of the models, you can reuse them, but ensure their values are comprehensive enough to cover both models. For instance, if Business Units A, B, and C exist in Semantic Model 1, but A, B, and D are found in Semantic Model 2, a separate dimension table will be required.
Keep in mind that DirectQuery limitations still apply in this setup. Row-Level Security (RLS) from either semantic model will still be enforced, and unlike traditional DirectQuery connections to databases, data transformations are not supported.
