Well, that’s what happens behind the scenes anyway. If there was a p5.js function to which you could pass CSV content as a string, that function would still have to break the string apart and build a p5.Table from those parts. It wouldn’t be a less expensive operation just because another programmer implemented the function for you.
I wouldn’t worry too much before you actually try it, though. Maybe it will go faster then you expect, and you only have to do it once at the beginning of your sketch, during setup.
Just because I’m curious, what’s the reason you can’t have the CSV in a file and use the built-in loadTable function?