httpd has asked for the wisdom of the Perl Monks concerning the following question:

Hi,

that's just a simple question, but I cannot find a way to operate with data in var, for example, what I can do using opened FILE Descriptor: 

open(query, "netstat -na | awk '{print \$4}' | sort -nr |") || die "fi
+le not found!";

 while(<query>)
{
($num, $ip) = split(' ');
print "output: $num $ip\n";
}
close(query);

[download]

How can I do the same with using VAR instead of FILE descriptor? with some thing like this: 

$query =  `netstat -ntu | awk '{print \$4}' | sort -nr`;
while(<$query>)
{
($num, $ip) = split(' ');
print "output: $num $ip\n";
}

[download]
However this "while" construction does not work, can you please help me? may be I should use "foreach" or something else?

Replies are listed 'Best First'.
Re: simple question about printing vars
by moritz (Cardinal) on Nov 03, 2011 at 12:45 UTC

    Backticks just return a string, no magic involved. You can't readline (which is what <$var> does) from a string.

    A possible solution is to open an in-memory file handle: 

    open my $handle, '<', \$query or die "Cannot open in-memory file handl
+e: $!";
while (<$handle>) { ... }

    [download]

    Or you can use the backticks in list context: 

    for (`...`) {

}

    [download]

    Though for this particular problem, the pipe open is still the best (and least memory hungry) solution.

    Perl 6 - second systems done right
[reply]
[d/l]
[select]
      thanks, I will use pipe then.
[reply]
Re: simple question about printing vars
by choroba (Cardinal) on Nov 03, 2011 at 12:43 UTC
    Use reference as the filename: 
    $query =  `netstat -ntu | awk '{print \$4}' | sort -nr`;
open my $FH, '<', \$query;
while (<$FH>) {
    ...
}

    [download]
    Also, consider doing the awk's work and sorting in Perl, too.
[reply]
[d/l]

      thanks, so as far as I understand using FILE descriptor is the best solution, right?

      as for awk, uniq and sort bash commands, what is the equivalent in perl for that functions, or should I use regexp?

      for example: I need to sort IPs from netstat command, look: 
      # netstat -ntu | egrep 'udp|tcp' | awk '{print $4}' | sort | uniq -c |
+ sort -nr
     17 24.13.201.116:80
      1 44.43.121.146:22

      [download]

      how can I do the same using just perl functions?

      Is it better to use perl native functions for such purpose instead of calling external shell commands?
[reply]
[d/l]
        Something like this?
        netstat -ntu | perl -nae 'next unless $F[0] =~ /tcp|udp/;
    $seen{$F[4]}++
} {
    @keys = sort { $seen{$b} <=> $seen{$a} } keys %seen;
    printf "% " . (1 + length($seen{$keys[0]})) . "d %s\n",
           $seen{$_}, $_
        for @keys
'

        [download]
[reply]
[d/l]
        Maybe like this? 
        #!/usr/bin/perl
use strict;
use warnings;

open PO, "netstat -na|" or die "Can't pipe open: $!\n";
  my %count;
  while (<PO>) {
    next unless /^udp/ or /^tcp/;
    my @fields = split;    # split $_ on whitespace
    $count{$fields[3]}++;
    ## next jumps here
  }
close PO;

for my $ip (sort keys %count) {
  print "\t$count{$ip} $ip\n";
}

        [download]

        Have a nice day, j

[reply]
[d/l]

Back to Seekers of Perl Wisdom