Summary

In this chapter, we began to dive deeper into specific areas of an actual Power BI solution, starting from transforming and loading data. We saw how Power Query and the mashup engine take center stage in this part of the pipeline, powered by the M query language. We learned how memory and CPU are important for data refresh operations. This meant that poor Power Query design can lead to failed or long-running data refreshes due to resource exhaustion.

Additionally, we learned about parallelism and how you can change the settings in Power BI Desktop to improve performance. There are also settings that can be adjusted in Power BI Desktop to speed up the developer experience and optimize data loading in general. We also learned how to customize refresh parallelism in Power BI Premium, Embedded, and Azure Analysis Services.

Then, we moved on to transformations, focusing on typical operations that can slow down with large volumes of data such as filtering, joining, and aggregating...