Use a CSV parser so that you don't have to cook your own regular expressions for a job that a full blown parser is better at. Then it should be easier to find the numeric fields, and if necessary reformat them to look more like a number you can work with. Your home made CSV parser is obviously choking on the "commas as data versus commas as field separators" issue. It's already been solved. Text::CSV::Simple is one example of a parser that could simplify the process for you.