in reply to edit a CSV and "in-place" replacement
Weighing in here, there is n-o .. w-a-y (... “in helll...”) that I would approach this problem in any way other than to first handle the record in a CSV-processing Perl package, then do the manipulation on the 35th item, then use the same package to reassemble the string.
The key issue here, of course, is maintainability. (Got CPU-cycles to burn ... don’t care one whit about that.) The solutions that you pick must not only work “right now,” but they also must be durable. Fixed in such a way that it stays fixed, forever. No matter what (within reason) you or somebody else does next. Otherwise, you wind up creating fragile code, full of so many functional interdependencies that it just can’t stand-up any more on its own two feet. Heh, it’s a hall-of-mirrors in which every single mirror is cracked. (Been there, seen that. All of us have, ’round these parts, and we’re still slurpin’ the Alka-Seltzer for it.) Not pretty.) No, when you do one thing to a complicated structure (and a CSV-file actually does qualify as “complicated,” as the authors of those CPAN modules can tell you), you want to be very sure that you’re taking the long view of things. Even though we do like to fool around with things like “golf” around here, just for our own entertainment, cleverness actually does not win Brownie Points.