I’ve worked through the book and it starts to click in my head. What’s missing for me are examples how to fallback to some lower level stuff like a custom ecto query with fragments. Sometimes some data needs to be fixed in bulk. It can’t be done in a single query, and the easiest way is to use ecto query in a recursive function that updates the data in smaller chunks in the background. It’s pretty simple using pure ecto, but I’m not sure how do do it using ash resources.
The good news is, you can always fall back to writing Ecto queries directly, as Ash resources are built on top of Ecto schemas. I tend to do this if I need to write data migrations - using resource actions there isn’t great because then the migrations need to be kept up to date (otherwise your app won’t compile).
I’ll have a think about this one, see if we can make a mention about it somewhere. Thanks for the feedback!