Hmmm that doesn't seem super fluent, but definitely doable. Thanks for the advice.
Have you had any experience with dbplyr? I'm not too familiar myself but I was hoping there might be some solution where you could handle the dataframe, and only instatiate it (chunkwise) as needed.
I've seen similar things in python ORMs (Django, Peewee). Although they tend to leverage python's generators, so I'm not sure how doable such a thing is in R