Forum Discussion
Single Slicer for shared dimensions across multiple Power BI datasets?
Hi MahmoudZou ,
Thanks for the clarification.
Just to add to what parry2k mentioned, this is expected with the current composite model behavior. Even if the business semantic models originate from the same shared dimensions dataset, Power BI still treats each semantic model as a separate source group. Because of that, the dimension tables do not automatically merge and the relationships defined inside the source models are not inherited when multiple semantic models are combined in one report.
As you noted, loading only the fact tables would require rebuilding relationships in the Balanced Scorecard, which defeats the purpose of reusing the existing logic. At the moment there is no supported feature that recognises two dimension tables from different semantic models as the same table and keeps a single slicer without extra modeling.
For more details, please visit:
Use composite models in Power BI Desktop - Power BI | Microsoft Learn
Composite model guidance in Power BI Desktop - Power BI | Microsoft LearnThere are similar community threads discussing this scenario:
Solved: Composite model relationships / shared dimension t... - Microsoft Fabric Community
Shared Dataset and Composite Dataset - Microsoft Fabric CommunityOne practical approach is to create a single semantic model that contains the shared dimensions once and then connects to the business facts. The Balanced Scorecard would then rely on that single model instead of combining multiple semantic models at the report layer, avoiding repeated relationship work.
You can monitor the Power BI release plans and Fabric blog for any future changes in this area.
Hope this helps. Please reach out for further assistance.
Thank you.
Super UserMahmoudZou, when you are connecting to the semantic model, you have the choice to pick the tables you want. Maybe you don't want to pick these dimension tables and use it from the common dimension table dataset, and then create the relationship in your new composite model.
Thank you for the suggestion. I understand the approach you're recommending, but if I only select the fact tables from the business datasets and omit their dimension tables, I lose all the pre-defined relationships and context from the source semantic models. Manually recreating the relationships for all the fact tables in the composite model contradicts the main objective. The goal is to reuse the existing business logic and relationships from the source datasets without any manual rework or duplication of effort and ensure any changes in the source models are automatcally reflected in the Balanced Scorecared.
Regards
Mz
v-veshwara-msftCommunity Support
Hi MahmoudZou ,
Thanks for the clarification.
Just to add to what parry2k mentioned, this is expected with the current composite model behavior. Even if the business semantic models originate from the same shared dimensions dataset, Power BI still treats each semantic model as a separate source group. Because of that, the dimension tables do not automatically merge and the relationships defined inside the source models are not inherited when multiple semantic models are combined in one report.
As you noted, loading only the fact tables would require rebuilding relationships in the Balanced Scorecard, which defeats the purpose of reusing the existing logic. At the moment there is no supported feature that recognises two dimension tables from different semantic models as the same table and keeps a single slicer without extra modeling.
For more details, please visit:
Use composite models in Power BI Desktop - Power BI | Microsoft Learn
Composite model guidance in Power BI Desktop - Power BI | Microsoft LearnThere are similar community threads discussing this scenario:
Solved: Composite model relationships / shared dimension t... - Microsoft Fabric Community
Shared Dataset and Composite Dataset - Microsoft Fabric CommunityOne practical approach is to create a single semantic model that contains the shared dimensions once and then connects to the business facts. The Balanced Scorecard would then rely on that single model instead of combining multiple semantic models at the report layer, avoiding repeated relationship work.
You can monitor the Power BI release plans and Fabric blog for any future changes in this area.
Hope this helps. Please reach out for further assistance.
Thank you.
Thanks v-veshwara-msft for sharing the links. I’ll give this option a try and see how it performs, especially in terms of efficiency.
- v-veshwara-msft
Thanks for the update MahmoudZou . If you face any issues while testing this approach, please share the details and we can review further.