Data Modelling in MS SQL vs in Power BI/DAX

[DepthOfFocus]

Hi All -

Our organization is moving our reporting from Excel over to Power BI.

Our source data comes from various sources (Oracle Tables, MS SQL Tables, CSV files, Analysis cubes). As such, it's always been a standard practice for us to consolidate and model our data using MS SQL server, to output basic flat tables to be easily consumed by Excel Pivot tables. 

As we move to power BI, is it a good/bad idea to continue to model our data using MS SQL? Or is it a better practice to entirely use Power BI/DAX to link/consume/model the data, and not use MS SQL as an intermediary?

thanks!

[jdbuchanan71 · Accepted solution]

@DepthOfFocus 

Most of my reading points towards using SQL to create views that are then fed to PowerBI where you create the model (relationships and measures).  This allows good control from an IT governance perspective.  Take a look at this article from the team at SQLBI.

https://www.sqlbi.com/articles/data-import-best-practices-in-power-bi/