I recently had a similar task of having to organize CSV files. I originally started out using DBD::CSV because I wanted the SQL like access to the information. However, I ended up using Text::CSV_XS. I did this for a number of reasons.

The first reason was that I was also able to use Text::CSV::Simple which allowed me an easy interface (since many of the CSV files I was working with had less than 15 columns in them). I also used it because it has a simple straightforward interface to handle quoted delimiters (ex. val1,val2,"val,3",val4) which is something I needed.

You weren't specific as to what you were using it for, but if you are just merging CSV files, some of the modules are very useful (at least they were for what I needed them for).