The basic process is still the same. You record what you have seen and make processing decisions based on that data.

Providing appropriate values for $key, doSomething(), and $data are left as an exercise for the reader.

Update: Perhaps I am misunderstanding your question. Please post what you expect to see as output given the provided input data.

Sorry still you didn't understand my problem I think, In the above input data as I provided, in that data number tag is same in first data node and third data node, but reaming information is different only number tag is same. I need to delete the entire node if number tag element is same in any node. I am printing $info into new file so before printing into new file I need to eliminate duplicate nodes by using number tag. I hope now you understand clearly, I posted my expected result also. can you suggest me.

Given your output data, the approach I provided will do the job. $key should be the 'number tag element'. Don't add the results if you have already seen the number.

This block of code illustrates the technique.

my @data = qw( a b 3 d a f b );
my %seen;
for ( @data ) {
    print $_, "\n" unless $seen{$_}++;
}

__DATA__
a
b
3
d
f
[download]

You should be able to apply this technique to your problem space.

--MidLifeXis