Appending tables = Improved performance?

[Anonymous]

Hi all,

First time poster and somewhat new with Power BI.  I'm currently throwing together a model that has several very similar tables.  One table contains all the details for products but for our eastern region, the other contains data for western, and both join to the central fact table.  I know appending these tables is generally best practice just for making the model simpler to follow, but does this also improve the performance in general?  I'd assume so as it's less joins to manage, but I'm just asking for my own basic knowledge.  It's worth noting it has the same columns in both tables.

Thanks in advance!

[edhans · Accepted solution]

Absolutely. Power BI performs best when you are using a star schema, so rather than multiple product tables, you'd want one product table with all of the relevant fields in it. This is very different from how you'd design a database, but databases are generally designed to optimize read/write/delete operations. Power BI is optimized for query operations, and duplicated data doesn't cause an issue. 

See this article, or google Power BI Star Schema for more info, or see this book that goes deep into data models and how they are best built for Power BI.

[Anonymous]

Thank you so much, edhans!