Handling Schema Changes in Import Mode – Dynamic Columns in Power BI Service

Hi. As a good practice knowing the columns you need for the report analysis is key. Let's consider this, it doesn't matter if the source has 7, 8 or 10 columns depending on the sql server you pick if you know the 5 columns you should use. Unless those 5 change, you won't care for more columns because even if they are added at the semantic model, you won't see them at the report. The visualizations are built with the 5 you need. There is no dynamic changing visual columns. The problem would be if those 5 necessary columns change because the visuals will be broken and you need that column for the analysis.

My recomendation is to analyze the sources and changes of the tables to make sure you can hold the columns you need for the dashboards. Otherwise it won't change if you can dynamically add them to semantic model or not. They will break anyway when you change a necessary column and won't care if it's a new one that is not used.

Consider that Power Query code is the key to dynamic operations. If you are using a step in the code asking for a specific column name, then you can't miss that column in the source. You need to find a way to reference everything without writing the name at the code if you want to keep it dynamic.

I hope that helps,