If you're merely trying to delete lines beginning with ';', then the following should work.

perl -ni.bak -e 'print unless /^;/' file
[download]

Thanks but i want to run it from within the program. I program does other things too. It is only part of the program, so i can't call the program like that. But thank it is still a very usefull hint you gave me.

my @data = grep { /^\S+;/ } <FILEHANDLE>;
[download]

Then the lines you want are in @data.

--
<http://www.dave.org.uk>

"The first rule of Perl club is you do not talk about Perl club."
-- Chip Salzenberg

If the "other things" in your program are all happening inside a while loop like this:

while (<>) {  # or maybe using a file handle, like <IN>
   # do stuff...
}
[download]

and you just want to skip lines that start with a semi-colon, then just add one line at the top of the loop like this:

while (<>) {
    next if ( /^\s*;/ );  # skip if first non-space char is ";"
    # now do stuff...
}
[download]

If you also want to delete lines with double semi-colons (because these represent empty fields), adjust that first line within the loop as follows:

while (<>) {
    next if ( /^\s*;/ or /;;/ );
    # now do stuff...
}
[download]

The "next" statement simply causes the script to skip everything that follows in the loop, and go right to the next iteration. The regular expressions in the "if" statement are saying:

    ^    -- line begins with...
    \s*  -- zero or more white-space characters
    ;    -- followed by a semi-colon
or
    ;;   -- there are two semi-colons together anywhere in the line
[download]