Forum Discussion

Manoshi's avatar
Manoshi
Frequent Visitor
11 months ago
Solved

Best Practice for Building and Sharing Semantic Models in Microsoft Fabric

Hi everyone, I’m looking for guidance on best practices for designing my architecture in Microsoft Fabric. My main goal is to create a semantic model using Fabric Warehouse data, build reports on to...
microsoft fabric
power bi
  • wardy912's avatar
    wardy912
    11 months ago

    Hi Manoshi , use onelake security for your RLS/CLS requirements. That way, you are applying the restrictions at the table level, and the restrictions will follow the data in any platform it is used, rather than being specific to the semantic model.

tayloramy's avatar
tayloramy
Icon for Super User rankSuper User
11 months ago

As others have mentioned, it really depends on what your goals are. 

Import mode is the best performance, but has memory limitations. DirectLake is the best balance between performance, maintenance, and being able to work with large datasets. 

I would avoid direct query against the SQL endpoint, that just adds a layer of latency that DirectLake doesn't have.