I apologize for leaping to the worst interpretation of your post! I hope the code below helps. It will muck up the date ordering, but I've run out of time tonight to clean that up.

use strict;
use warnings;
use Data::Dump::Streamer;

my @data = (
    ['01 Jul 2007','221.6.19.196',9,19],
    ['01 Jul 2007','221.6.19.197',1,3],
    ['01 Jul 2007','221.6.19.196',12,12],
    ['01 Jul 2007','221.6.19.197',2,2],
    ['02 Jul 2007','202.119.104.22',1,1],
    ['02 Jul 2007','221.6.19.196',20,45],
    ['03 Jul 2007','202.119.104.12',1,11],
    ['03 Jul 2007','202.119.110.236',1,2],
    ['03 Jul 2007','210.29.132.9',1,2],
    ['03 Jul 2007','210.29.141.188',1,2],
    ['03 Jul 2007','221.6.19.195',3,7],
    ['03 Jul 2007','221.6.19.196',4,7],
    ['03 Jul 2007','222.192.2.213',1,0],
    ['03 Jul 2007','202.119.108.42',3,3],
    );
my %sums;

$sums{$_->[0]}{a} += $_->[1], $sums{$_->[0]}{b} += $_->[2] for 
    map {["$_->[0],$_->[1]", @{$_}[2,3]]} @data;
    
my @result = map {[@{$_}[1, 2, 3, 4]]}
    sort {$a->[0] cmp $b->[0]}
    map {[$_, split (',', $_), @{$sums{$_}}{'a', 'b'}]} keys %sums;

Dump (\@result);
[download]

Prints:

$ARRAY1 = [
            [
              '01 Jul 2007',
              '221.6.19.196',
              21,
              31
            ],
            [
              '01 Jul 2007',
              '221.6.19.197',
              3,
              5
            ],
            [
              '02 Jul 2007',
              '202.119.104.22',
              ( 1 ) x 2
            ],
            [
              '02 Jul 2007',
              '221.6.19.196',
              20,
              45
            ],
            [
              '03 Jul 2007',
              '202.119.104.12',
              1,
              11
            ],
            [
              '03 Jul 2007',
              '202.119.108.42',
              ( 3 ) x 2
            ],
            [
              '03 Jul 2007',
              '202.119.110.236',
              1,
              2
            ],
            [
              '03 Jul 2007',
              '210.29.132.9',
              1,
              2
            ],
            [
              '03 Jul 2007',
              '210.29.141.188',
              1,
              2
            ],
            [
              '03 Jul 2007',
              '221.6.19.195',
              3,
              7
            ],
            [
              '03 Jul 2007',
              '221.6.19.196',
              4,
              7
            ],
            [
              '03 Jul 2007',
              '222.192.2.213',
              1,
              0
            ]
          ];
[download]

Thanks a bunch GrandFather, I am working on your solution to fit my requirment. Sorry in sending the delayed response. Will let you know once done. Cheers, Dinakar

It occurred to me on the walk to work this morning that the sort problem can be fixed by introducing a sort key:

# @data as used previously
my %sums;
my $sortKey;

for my $datum (map {[++$sortKey, "$_->[0],$_->[1]", @$_]} @data) {
    if (exists $sums{$datum->[1]}) {
        $sums{$datum->[1]}[4] += $datum->[4];
        $sums{$datum->[1]}[5] += $datum->[5];
    } else {
        $sums{$datum->[1]} = $datum;
    }
}
   
my @result = map {[@{$_}[2 .. 5]]} sort {$a->[0] cmp $b->[0]} values %
+sums;

Dump (\@result);
[download]

Output is as shown previously - the dates are all in the same month.

---

DWIM is Perl's answer to Gödel

Thanks a lot GrandFather, It worked well for me :) You have made my day easier. thanks, ~Din

Sure:

for my $datum (map {[++$sortKey, "$_->[0],$_->[1]", @$_]} @data) {
    if (exists $sums{$datum->[1]}) {
        $sums{$datum->[1]}[$_] += $datum->[$_] for 4 .. $#$datum;
    } else {
        $sums{$datum->[1]} = $datum;
    }
}
   
my @result = map {[@{$_}[2 .. $#$_]]} sort {$a->[0] cmp $b->[0]} value
+s %sums;
[download]

Take note of the two places where $# is used to get the last array index. Note that this technique deals with differing column counts on a row by row basis. Unlikely to be an issue in your current case I'd guess, but neat anyway. ;)

The code does assume that columns are left aligned and thus, any missing columns are assumed to be missing from the high index end of the array (in my mind index 0 is left of index 1).

---

DWIM is Perl's answer to Gödel

If this is a table why don't you let the database do the work?