If I understand you correctly, you want

• Step through a text file line by line
• If you come across a line with a certain set of characteristics ...
• Then you want to do something to the line just prior

You have a number of options. The easiest to implement is to read the file into an array, then manipulate it. Something like this:

my @lines = <>;
foreach my $index (0 .. $#lines) {
    if ($lines[$index] =~ /word/) {
        # Do what you want here, but to $lines[$index - 1]
    }
}
[download]

The only problem with that method is if your text file is over 50% of your available RAM. Then, instead of being blazingly fast, it's only moderately fast.

Originally posted as a Categorized Answer.