Hashes are very useful for matching.

use strict;
use warnings;

my %lookup;

{
   open(my $file1, '<', 'file1.txt') or die("...");

   while (<$file1>) {
      chomp;
      my @fields = split(/\|/, $_);
      $lookup{$fields[0]}{$fields[1]} = $fields[2];
   }
}

{
   open(my $file2, '<', 'file2.txt') or die("...");

   while (<$file2>) {
      chomp;
      my @fields = split(/\|/, $_);

      if (exists $lookup{$fields[0]}{$fields[1]}) {
         $fields[3] = $lookup{$fields[0]}{$fields[1]};
      } else {
         $fields[3] = 'NO MATCH';
      }

      print(join('|', @fields), $/);
   }
}
[download]

Tested.

# $f1 - small file
# $f2 - big file

open my $fh, $f1 or die "cannot open $f1: $!\n";

my @file1 = <$fh>;
my %file1 = map{ chomp and
                    (join '|', (split /\|/)[0,1]),
                    (join '|', (split /\|/)[2])
            }@file1;

open my $fh, $f2 or die "cannot open $f2: $!\n";

while(<$fh>){
        chomp;
        my $part = (join '|', (split /\|/)[0,1]);
        print $_, '|',
                (exists $file1{$part} ? $file1{$part} : 'NO MATCH' ), 
+$/;
}
[download]

sort f1 | perl -pe 's/\|/~~/' > f1~
sort f2 | perl -pe 's/\|/~~/' > f2~
join -t'|' -a 1 -a 2 f2~ f1~ \
         | perl -pe 's/~~/|/' \
         | perl -F'/\|/' -ape 's/$/|NO MATCH/ unless $#F==3; $_'
[download]

  SELECT f2.A, f2.B, f2.C, COALESCE(f1.C, 'NO MATCH') as D
  FROM f2 LEFT JOIN f1 ON f1.A = f2.A AND f1.B = f2.B
[download]

Hashes are good for this kind of thing:

use strict;
use warnings;
use Carp;

my $hr_f1 = getlines($ARGV[0]);

open(IN,"<$ARGV[1]") or croak "unable to open $ARGV[1]: $!";

while(my $l = <IN>) {
        my ($k,$v) = getkv($l);
        if (defined $hr_f1->{$k}) {
                print "$k|$v|$hr_f1->{$k}\n";
        } else {
                print "$k|$v|NO MATCH\n";
        }
}

sub getlines {
        my $fn = shift;
        my %h = ();
        open(IN,"<$fn") or croak "unable to open $fn: $!";
        while(my $l = <IN>) {
                my ($k,$v) = getkv($l);
                $h{$k} = $v;
        }
        close(IN);
        return \%h;
}

sub getkv {
        my $l = shift;
        chomp $l;
        my @l = split /\|/, $l;
        return "$l[0]|$l[1]","$l[2]";
}
[download]

and tested:

$ ./join.pl f1 f2
789|efg|2222222|NO MATCH
123|abc|9999999|777
786|uvw|1234567|NO MATCH
123|xxx|0000000|NO MATCH
234|cde|0000000|456
567|xyz|1111111|999
[download]

Update: This is the same as ikegami's solution...guess i wasn't quick enough...

#!/usr/bin/perl

use strict;
use warnings;


# substituted for the FILE
my @file2 = qw(789|efg|2222222 123|abc|9999999 786|uvw|1234567 123|xxx
+|0000000 234|cde|0000000 567|xyz|1111111);


# process file 1

my %file_1_hash = ();

while ( <DATA> ) {
  chomp;
  my @array = split( /\|/ );
  $file_1_hash{join( '|', @array[0..$#array-1] )} = $array[$#array];
}

# when file2 is a FILE, just do a while loop here

foreach (@file2) {
  my @file_2_array = split( /\|/ );
  my $file_2_key = join( '|', @file_2_array[0..$#file_2_array-1] );

  if ( !exists $file_1_hash{$file_2_key} ) {
    $file_1_hash{$file_2_key} = "NO MATCH";
  }

  print join( '|', $file_2_key, $file_2_array[$#file_2_array], $file_1
+_hash{$file_2_key} ), "\n";
}

__DATA__
123|abc|777
234|cde|456
567|xyz|999
[download]