G'day vkk05,

"... looking for any alternative solution/logic ..."

The code in &get_max_count below provides a far less involved solution.

Testing your algorithm with a single, ideal datum is never a good idea. I've used Test::More to show a common way of testing multiple items. You can simply add more test data to @tests; nothing else in the code needs to be changed.

#!/usr/bin/env perl

use strict;
use warnings;
use constant UNDEF => "\0";

use Test::More;

my @tests = (
    [qw{2 2 2 3 3 4 4}],
    [qw{2 2 3 3 4 4 2 2 2}],
    [qw{2 2 2 3 3 4 4 2 2}],
    [qw{2 2 2 3 3 4 4 4 5 6 6}],
    [qw{2 2 6 6 6 3 3 4 4 4 5}],
    [qw{2 2 3 3 4 4}, ('') x 3],
    [('') x 3, qw{2 2 3 3 4 4}],
    [('') x 2, qw{2 2 3 3 4 4}, ('') x 3],
    [qw{2 2 3 3 4 4}, (undef) x 3],
    [(undef) x 3, qw{2 2 3 3 4 4}],
    [(undef) x 2, qw{2 2 3 3 4 4}, (undef) x 3],
    [(undef) x 2, ('') x 2, qw{2 2 3 3 4 4}, ('') x 2, (undef) x 3],
);

plan tests => 0+@tests;

ok(get_max_count($_) == 3) for @tests;

sub get_max_count {
    my ($data) = @_;
    my ($inst, $key, %consec, $last);

    for (@$data) {
        $_ = UNDEF unless defined;
        $last = "$_:$_" unless 0+keys(%consec);

        if ($_ ne $last) {
            $key = join '-', $_, ++$inst;
            $last = $_;
        }

        ++$consec{$key};
    }

    return (sort { $b <=> $a } values %consec)[0];
}
[download]

Output:

1..12
ok 1
ok 2
ok 3
ok 4
ok 5
ok 6
ok 7
ok 8
ok 9
ok 10
ok 11
ok 12
[download]