In this article I will not be looking at Power BI as a data visualization tool. Indeed, I will not be discussing in anything more than a cursory way the data transformation and modelling capabilities of this product. Instead, I would like to focus on using Power BI Desktop to connect directly to certain data sources. So although you need no previous knowledge of Power BI Desktop to understand this article, I will be providing only an extremely succinct glance at the extensive array of the data mashup and modelling techniques that are available. Nor will I be explaining any of the approaches that can be taken to create visuals and dashboards. I will, however, try out a few classic techniques using the Power BI Query Editor and Data Model to see if there are any limits to using DirectQuery when compared to data loaded into the in-memory model. The aim here is to see just how, and in what circumstances, DirectQuery can help you to analyze your data faster and more easily.
This article is lengthy but good. Check it out.