Troubles with grouping in Power Query

You have 2 options then I reckon. If it's a fabric pipeline, Spark based pipeline or an Azure Data Factory pipeline, you can add transformation steps to do the aggregations and then the Power Query engine won't need to spend time doing that effort.

Alternatively, the better option in my opinion, would be to make a view based on that table, within the SQL Database that you're querying. You can do things like grouping and roll up within SQL statements. If you're not the best at SQL then maybe use ChatGPT to try and convert your m code into SQL code, which you can use to make the view.