Hi all,
(If anyones still listening, as it were..)
Just out of interest, at the moment I am using demerphqs solution in my program. Thats not to say that any others are bad/worse/whatever, but I prefer one that I understand, and can change if necessary, most of those with 'cryptical' regexes I don't as much.
I tried correcting some errors in mine, to test for faulty input, and it wasn't that easy :) And since I wanted to see some results.. Maybe I'll fix it one of these days.

The difference in performance is noticable however, especially when I was importing a 500k file, using my functions imported it in about the time it took me to switch apps and back, using P::RD gives me enough time to make a cup of tea (or two) and probably a sandwich to go with it :) - Luckily it was a one-time import, parsing smaller inputs on the fly is fast enough not to notice the difference.

One small niggle: Some of the strings received contain escaped apostrophes  "Blah \'s Mud" which the parser returns as:  "Blah \\\'s Mud", seems to be adding \ instead of removing them :)
I had to add a cheat to mine which reconcatenates strings that had commas in them prior to me splitting them apart, so now Im looking for a better approach.. Maybe instead of splitting, attempting to chop off each type from the input string.. ?

Anyfish, I'll get back to you.. (And try out barts solution..)

C.