If i understood your question correctly, add e option modifier, to eavluate

$line = '<:KEYWORD:4:>';
$value = 5;
$line =~ s/<:KEYWORD:([^:>]*):?>/($1 != 0)? ($value * $1) : $value/e;
print "$line";
[download]

updated:

duh!!
*slaps forehead*
I forgot all about the 'e' modifier... I really should pay closer attention to the doco.

Thanks,
--Darren

Hai, Try this,

use strict;
my $value=5;
while (my $line=<DATA>){
$line =~ s/<:KEYWORD:([^:>]*)\:?>/my $dum=$1; my $val; if ($dum) { $va
+l = $value * $dum }else{ $val=$value} $val/e;
print $line, "\n";
}
__DATA__
<:KEYWORD:4:>
<:KEYWORD:5:>
<:KEYWORD:>
<:KEYWORD:9:>
[download]

You can rewrite

my $dum=$1; my $val; if ($dum) { $val = $value * $dum }else{ $val=$val
+ue} $val
[download]

To be ($1 != 0)? ($value * $1) : $value like prasadbabu had it, or really it can just be boiled down to:

$value * ($1 || 1)
[download]

(Assuming that $value is alway numeric) No need for a temp variable or the explicit if/else.

also, since you're not chomp'ing $line, that "\n" in your print line is going to going to double-space the output...

Thanks :)

This is a nice, clean option and is what I've settled on.


$line =~ s/<:KEYWORD:([^:>]*):?>/int($value * ($1 || 1))/e;
[download]

The not chomp'ing was never an issue, as I wasn't explicity adding newlines to the output.

heh... I've just been browsing through my copy of Perl Best Practices, and realised that I should abstract my KEYWORD<=>$values pairs into a lookup hash.... so I'm off to do that now :)

Thanks again,
--Darren

use strict;
use warnings;
my %values = ( foo => 3, bar => 4, stuff => 5 );
while( <DATA> ){
  s/<:(KEYWORD):(?:([^:>]+:)?>/ $values{$1} * ($2 || 1) /e;
  print;
}
__DATA__
<:foo:4:>
<:bar:5:>
<:bar:>
<:stuff:9:>

# OUTPUT:
12
20
4
45
[download]

McDarren - I realise you've got most of this working already. You're saying that you're trying to abstract the keyword-lookup. I've done something very similar, and posted it here. In your case, you want to change the delimiters, but I think it can be used as a framework for a very good generic-lookup-and-then-modify tool. In fact, I use this code in production.

Just create an expand_var function, and use that to do lookups.

#!/usr/bin/perl -w
                                                                      
+          
use warnings;
use strict;
no strict 'refs';
                                                                      
+          
our $a = 10;
our $b = 20;
our $c = 33;
our $d = 123;
                                                                      
+          
my @lines = qw( <:a:0:> <:a:3:> <:b:0:> <:b:4:> <:c:5:> <:d:6:> );
                                                                      
+          
foreach my $line (@lines) {
    if ($line =~ s/<:(\w+):(\d+):>/$$1*($2||1)/e) {
        printf "Result = %d\n", $line;
    }
}
[download]

    Result = 10
    Result = 30
    Result = 20
    Result = 80
    Result = 165
    Result = 738

heh... "substituting certain keywords with variables"... was badly worded - what I should have said was: "substituting certain keywords with the values of associated variables"

For example, I have stuff like:

    $line =~ s/<:INCOMING:([^:>]*):?>/int($incoming * ($1 || 1))/e;
    $line =~ s/<:OUTGOING:([^:>]*):?>/int($outgoing * ($1 || 1))/e;
    $line =~ s/<:INGRESS:>/$eth1/;
    $line =~ s/<:EGRESS:>/$eth0/;
    if ($line =~ /<:RADIUS:>/) {
        foreach my $ip (@radius_ips) {
            $thisline = $line;
            $thisline =~ s/<:RADIUS:>/$ip/;
            print OUT $thisline;
        }
    }
[download]

(I'm in the process of abstracting the keyword/value pairs into a lookup hash, as per my previous post)