probably don't need perl. there's a shell command that can do this.

% man sort

so the whole thing might go something like:

sort -n /etc/hosts > /tmp/hosts.sorted
mv /tmp/hosts.sorted /etc/hosts
[download]

anders pearson

If you must do it in Perl use the Schwartzian Transform for efficiency. This avoids finding the 4th octet more than once for each line and is thus fast(er) than it might otherwise be.

my @unsorted = <DATA>;

my @sorted =
  map{ $_->[0] } 
  sort{ $a->[1] <=> $b->[1] } 
  map { [$_, get_fourth($_) ] } 
  @unsorted;

sub get_fourth {
    local $_ = shift; 
    /\d+\.\d+\.\d+\.(\d+)/;
  return $1; 
}

print "@sorted";

__DATA__
10.194.196.55 heckyll
10.194.196.57 txwl013w
10.194.196.61 krios
10.194.196.62 stasis # dial up server
10.194.196.63 nnocdts01 # sparc 5 in wan rm
10.194.196.65 libra
10.194.196.66 legato-school
10.194.196.67 txwlxdbs nnocs002a
10.194.196.46 chqnocp01 # Laser 4MV by Joe Blow
10.194.196.47 chqnocp02
10.194.196.49 nnocrpt01
10.194.196.64 saturn
10.194.196.50 nnocgw001
10.194.196.52 nnocs002
10.194.196.53 wslkms02 wslknms0
10.194.196.54 nnocws001
[download]

Update

Changed order to ascending numrical order as per spec (I think)

cheers

tachyon

s&&rsenoyhcatreve&&&s&n.+t&"$'$`$\"$\&"&ee&&y&srve&&d&&print

sort -t. +3n /etc/hosts
[download]

excuse me I replied to the wrong reply. yes

sort -t +3n /etc/hosts
[download]

works !

thx again

-mike

Yes, I tried a sort -n but, here is a snippet of my output.

<code> 10.194.196.10 romulus
10.194.196.100 nnocws014
10.194.196.101 wslkms01 wslknms01
10.194.196.102 nnocws016
10.194.196.103 nnocws017
10.194.196.104 txwl008w
10.194.196.105 nnocws019
10.194.196.106 txwl015w
10.194.196.11 otter
10.194.196.110 nnocweb01
10.194.196.111 shemp
10.194.196.113 dvwl000s
10.194.196.125 spanky
10.194.196.126 buckwheat
10.194.196.127 datapro
10.194.196.128 darla
10.194.196.129 txwl018w
10.194.196.13 nnocdev01
<code>

this is not what we are looking for.

kind regards

-mike

updated to compensate for my foolish assumption.

use strict;
my @lines=<DATA>;
print sort{by_ip($a)<=>by_ip($b)} @lines;
sub by_ip {
    @_[0]=~ /\.(\d+)\s/;
    return $1;
}
__DATA__
10.194.196.55 heckyll
10.194.196.57 txwl013w
10.194.196.61 krios
10.194.196.62 stasis # dial up server
10.194.196.63 nnocdts01 # sparc 5 in wan rm
10.194.196.65 libra
10.194.196.66 legato-school
10.194.196.67 txwlxdbs nnocs002a
10.194.196.46 chqnocp01 # Laser 4MV by Joe Blow
10.194.196.47 chqnocp02
10.194.196.9 TEST
10.194.196.49 nnocrpt01
10.194.196.64 saturn
10.194.196.50 nnocgw001
10.194.196.52 nnocs002
10.194.196.53 wslkms02 wslknms0
10.194.196.54 nnocws001
[download]

Your example sorts lexically, not numerically. It'll sort 10.10.10.10 before 10.10.10.2, for example:

my @lines=<DATA>;
print sort @lines;
__DATA__
10.194.196.9  made-up
10.194.196.55 heckyll
10.194.196.57 txwl013w
10.194.196.61 krios
10.194.196.62 stasis # dial up server
[download]

This sorts the "made-up" line to the end, whereas the original poster presumably would want it first.

Like the other replies said, a numerical sort is more appropriate here. The number of "cycles" used either way is going to be pretty trivial.

add to your DATA: 10.194.196.9 nnocws001 10.194.196.100 nnocws001 and you'll see it's not quite so straight-forward. you're doing a sort based on "standard string comparison order", instead of numerical. though the OP doesn't specify.