Here's an example showing what you want to do with DBD::CSV:

#!/usr/bin/perl -w

#Define the behaviour of DBI->connect for the entire script
$ENV{DBI_DSN} ||= 'DBI:CSV(RaiseError=>1,AutoCommit=>1):'. join ';',
                  "csv_eol=\n",
                  'csv_sep_char=|',
                  'csv_quote_char=',
                  'csv_escape_char=';
     
use strict;
use DBI; 
use CGI;
  
my $statement = q{
  UPDATE list
     SET title = ?
       , description = ?
   WHERE id = ?
};
  
my $cgi = CGI->new;
  
#Connect to the data source
my $dbh  = DBI->connect;

$dbh->do(
  $statement, 
  {},
  $cgi->param('title'), 
  $cgi->param('description'),
  $cgi->param('id'),
);

$dbh->disconnect; 

print $cgi->header, 'Updated List';

__END__
[download]

Note: The above code assumes that your file name is "list". DBD::CSV can only process files without any extension. Also, if this is going into production you might want turn on Taint checking by appending a T to the first line, as well as validating each of the CGI paramters, possibly using HTML::FormValidator, which is my personal favorite for this.

The great thing about using DBD::CSV is that if you ever switch to using a relational database all the code won't need to change, assuming your schema stays the same. This is why I explicitly define the environment variable $ENV{DBI_DSN} at the top of the script, so that you can easily change it later. Keep in mind this varible globally defines the data source that all the DBI->connect method calls will use, and can be globally set for all your CGI's in Apache's httpd.conf.

As well, I do suggest you look into relational databases, like MySQL or PostgreSQL. They might be better suited for storing this type of information than a flat file.