Forum Discussion
Best practice - related columns in Power Query
You have correctly identified the 2 options. There is no 'best practice' in this scenario - whatever gets you to the model that you want is the correct way to go. Powerbi bringing back the related tables is a feature that you may or may not want to use. It can be handy, especially if you want to make use of the existing relationships in the source without digging around in the source db. However if you have full knowledge of the source data and you know how you want to transform data into the powerbi model, you may want to delete the related table columns and import them separately.
--
As long as you use a star schema and identify the Fact and Dimension tables you'll be on the right track. The way you get there is up to you.
My opinion is that it's not really about the relationships. It's about getting the data into powerbi.
SQL databases are usually normalised and optimised for storage based on selects/ inserts/updates/deletes i.e. the relational model. For powerbi, you will want a dimensional model (star schema) because it's purely about selects.
--
So it's not about a larger model, it's about the right model.