use strict;
use warnings;

use File::ReadBackwards ;

my $bw = File::ReadBackwards->new( 'file' ) or
    die "can't read 'log_file' $!" ;

my @lines;

$lines[0] = $bw->readline;

if ( $lines[0] =~ /pattern/ )
{
    for (1..9)
    {
        $lines[$_] = $bw->readline;
    }
}

@lines = reverse @lines;
        
print "@lines";
[download]

use Tie::File;

my @array;
tie @array, 'Tie::File', filename or die ...;
if ( $array[-1] =~ /regex here/ ) {
    # print elemeents -10 to -1.
}
[download]

File::ReadBackwards is significantly faster in this scenario.
Read more... (3 kB)
That can easily be explained. Tie::File always reads the whole file, File::ReadBackwards only reads the neccessary data.


holli, /regexed monk/

#!/usr/bin/perl -w 

# usefull on a unix-like system
# this uses tail - displays the last lines in a file
# 'tail -n 4 filename' shows the last 4 lines in a file 

my $file_to_check = '/tmp/test_file.txt';
my $file_to_store = '/tmp/fill_this_file.txt';
my $pattern = 'special text';

while (1)
{
     my $last_line = `tail -n 1 $file_to_check`;
    if ($last_line =~ /$pattern/)
    {
        `tail -n 11 $file_to_check >> $file_to_store`;
    };
    sleep(30 * 60);
}
[download]

Here's one way. Note that my example clobbers any existing output file and replaces that data with the new 10 lines. If you need to append the data to existing content, make sure to change that part of the script.

One thing I noticed is how old fashioned I am when it comes to array indices; I prefer using $lines[($#lines - 10) .. $#lines] rather than $lines[-11 .. -1]. I guess I just don't like a negative index? I'm strange that way :)

#!/usr/bin/perl -w

use strict;

my $in_file = 'in.dat';
my $out_file = 'out.dat';
my $spec_text = qr/special text here/;

while (1) {
    open(my $fh_in, '<', $in_file)
        or die("open failed: $!");
    my @lines = <$fh_in>;
    close($fh_in);

    if ($lines[$#lines] =~ $spec_text) {
        open(my $fh_out, '>', $out_file)
            or die("open failed: $!");
        print $fh_out @lines[($#lines - 10) .. $#lines];
        close($fh_out);
    }

    sleep 60 * 30; # 30 minutes
}
[download]