Forum Discussion
Semantic models per domain
Hi,
Now that we are migrating to Fabric, I’ve read about creating separate semantic models per domain, as well as separate datasets for shared dimensions.
In that setup, when you want to combine datasets, you would need to create a second-layer semantic model that uses DirectQuery instead of Import or DirectLake — is that correct?
Wouldn’t this approach introduce performance issues?
I’m very interested in hearing about any practical experiences or best practices regarding this.
Best regards,
Henri
Hi Henri,
Domains in Fabric is more related to (enterprise) organizations using data mesh architecture, where different department have their own (citizens) data developers and created their own extraction processes and models. In this architecture conformed dimensions need to exists so that all domains use the same dimensions and dimensions keys, so that data from different are interchangable.
But when you want to combine for example sales and procurement data, which could be seen as data from different domains, you still need to create a single semantic model to combine that data for optimized reporting.
Hope this helps. If so, please give a Kudos 👍 and mark as Accepted Solution ✔️.
2 Replies
Hi Henri,
Domains in Fabric is more related to (enterprise) organizations using data mesh architecture, where different department have their own (citizens) data developers and created their own extraction processes and models. In this architecture conformed dimensions need to exists so that all domains use the same dimensions and dimensions keys, so that data from different are interchangable.
But when you want to combine for example sales and procurement data, which could be seen as data from different domains, you still need to create a single semantic model to combine that data for optimized reporting.
Hope this helps. If so, please give a Kudos 👍 and mark as Accepted Solution ✔️.
Hi Hgroenevelt ,
May I know if your issue is resolved, or if you need any further clarification?
Please review the response shared by nielsvdc , and if you need any additional details, feel free to let us know.Thank you for your valuable update nielsvdc .
