Despite it not at all being the right tool for most jobs, this is why I actually enjoy writing spark. It reads top down, and is “sql-ish” for most commands. Enough that you don’t need to go out of your way to learn it if you already know SQL, you just think in terms of serial transformations.