My code was just a proof-of-concept to show how to use the range operator for this task. For a working script operating on files “log1.log”, “logX.log”, and “logZ.log”, say, — and assuming you want all the output to go to a single file “output.txt” (do you?) — you would do (untested):

my @filenames = qw(log1.log logX.log logZ.log);

open(my $out, '>', 'output.txt') or die "Cannot open file 'output.txt'
+ for writing: $!";

for (my $filename (@filenames)
{
    open(my $in, '<', $filename) or die "Cannot open file '$filename' 
+for reading: $!";

    while (<$in>)
    {
        ...

            if (/AS \s+ $keyword/ix)
            {
                print $out join('', @block);
                last;
            }

        ...
    }

    close $in or die "Cannot close file '$filename': $!";
}

close $out or die "Cannot close file 'output.txt: $!";
[download]

See perlintro and perlopentut.

Hope that helps,

So instead of opening an output file, and instead of looping over a list of files, do this:

while (<>) {
    do_something_here();
    print $blah if $accepted;
}
[download]

Then you run it like so:

my_script blah*.log > my_output_file
[download]

One of the things perl does really well is I/O, that is - it's really good at reading files and extracting data from them. It does it so well that speed is not even a consideration, in most cases. :-)

Like Athanasius said, you can keep of all that log content in their own files and read them from perl code. Then, you can use the Flip-flop operator as shown by Athanasius, reworking them to your needs. For starters, look at open and work from there. :-)