Rather than just using a hash, you want to use a hash of arrays, because you want to associate each key with multiple values. In this case, you might even want to use an Array of Arrays, since your keys are numeric. See HASHES OF ARRAYS in perldsc, and maybe perllol, perldata, perlreftut, and/or perlref for additional info.

As this signature is one of the best piece of advice in this field: is there a more complete collection than this newsgroup posting, apparently from dominus - perhaps even a canonical one?

My Google skills are as extensive as yours. When I adopted the quote, I did a lot of digging to see if there were longer versions of the list, and found none.

---

#11929 First ask yourself `How would I do this without a computer?' Then have the computer do it the same way.

For your output, you may also want to look at join and keys. A HoH or HoA could work for collecting the data for your results.

You may also want to put <code>...</code> around both your input and output data in order to preserve your intended formatting.

You might want to take kennethk's last bit of advice first. Work out the mechanism first by hand. That will inform your choice of control structure and data structure.

c:\@Work\Perl\monks>perl -wMstrict -MData::Dump -le
"my $pair = qr{ (\d+) \s+ ([[:alpha:]]+) }xms;
 ;;
 my $record = '1 A 1 B 2 G 2 H 2 V';
 ;;
 my %hash;
 while ($record =~ m{ \G \s* $pair }xmsg) {
   push @{ $hash{$1} }, $2;
   }
 dd \%hash;
 ;;
 $hash{$_} = join ':', @{$hash{$_}} for keys %hash;
 dd \%hash;
"
{ 1 => ["A", "B"], 2 => ["G", "H", "V"] }

{ 1 => "A:B", 2 => "G:H:V" }
[download]

$ perl -wMstrict -MData::Dumper -e 'my $pair = qr{ (\d+) \s+ ([[:alpha
+:]]+) }xms;
> my $record = "1 A 1 B 2 G 2 H 2 V";
> my %hash;
> while ($record =~ m{ \G \s* $pair }xmsg) {
>    $hash{$1} .= "$2;";
> }
> print Dumper \%hash;
> '
$VAR1 = {
          '1' => 'A;B;',
          '2' => 'G;H;V;'
        };
[download]

Why go to the trouble of using a hash?

#!/usr/bin/perl

# http://perlmonks.org/?node_id=1146923

use strict;
use warnings;

$_ = join "\t", qw(1 A 1 B 2 G 2 H 2 V);

1 while s/\b((\d+)\s\S+)(.*?)\s\2\s(\S+)/$1:$4$3/;

print "$_\n";
[download]

Thank you very much, can you explain me the regex in details , like what is \2 and $1 $4 $3 special variables purpose regards, Umesh

Hello bsherkhane,

The escape sequences \1, \2, \3, etc., are backreferences to captures in the current regex. The special variables $1, $2, $3, etc., are likewise backreferences to the captures in the most recent regex. $1 refers to the first capture, $2 to the second capture, and so on. Captures are numbered by counting left parentheses from the left. See perlre#Capture-groups.

The module YAPE::Regex::Explain is a useful tool for understanding regular expressions. Here is the explanation it gives for the left-hand side (i.e., the regex part) of the substitution in question:

Read more... (4 kB)

Hope that helps,

Athanasius <°(((>< contra mundum

Iustus alius egestas vitae, eros Piratica,