"Perl : Add an Array Element from one Array into another Array"

Too much questions for one question... But here we go anyway:

push @{$seen{$key}}, $item;

%seen is a hash (or associative array), so $seen{$key} recovers from %seen the value associated to that key $key. Then this value is treated as an array reference and transformed into an array by using @{} operator. Finally $item is added at the end of this array.

I don't understand what you mean by length... You mean the previous array length?

And to save this in a file, you just need to print() in your script and redirect to a file when executing the script, for example:

./my_perl_script.pl > my_output_file

Same goes for file input, you don't really need to open(), close() and such. This is more flexible and faster to code:

./my_perl_script.pl < my_input_file

This allows you to pipe this in a easier way and pass data from/to other scripts/processes. Of course both redirections can be used at the same time:

./my_perl_script.pl < my_input_file > my_output_file

Also you don't even need to save into a file (well it's always wise to have a copy of processed data anyway) and you can pipe directly the results to the other process, like

./my_perl_script.pl | my_other_script

This works on all OS I've used to far (Windows, Linux, OS X, BSD).

This

 $data = join ('',@data); #Convert to string
@data2 = split('\n',$data); #Explode along newlines 

Does not create your array like you think. It just recreates the line sructure you started with. I think you meant to split on "," commas. Use debugging tools. At a minimum insert a print block like this

print join(":", @data2); 

to see what is actually in your array.

Get each line working before you move on to the next one. Then, if you can't figure out why a line isn't working you can post a question here.

As it is, it's hard for use to tell what you were trying to say in the code because the ideas are incomplete.

UPDATE: I leave the link to the original answer's code to illustrate the concept of abstracting different subtasks (especially the processing one). But your problem can be solved much easier, if you're sure about what to expect in the input files:

use warnings;
use strict;

my $lengths_filename = 'Transcriptlength.fa';
my $counts_filename  = 'Quadcount.AllGtranscripts.fa';

my %sequence;   # it will be the basic data repository

local $/ = ''; 
# ...by this we ensure that files will be read by logical blocks instead of lines. 
# Might need some tweaking, if 'empty line' in your sample is not really empty.

# we start processing from 'counts' file, as only those records present in it
# should actually be in our output:
open my $cfh, '<', $counts_filename 
  or die $!, "\n";
while (<$cfh>) {
  # each logical block consists of two parts, divided by whitespace
  my ($name, $count) = split; 

  # here goes magic: we simultaneously create a new record in our repository...
  # ... and set its 'count' property to the value, extracted from the scanned fileblock
  $sequence{$name}{count} = $count;
}
close $cfh;

# now we go for lengths, approach is almost the same
open my $lfh, '<', $lengths_filename or die $!, "\n";
while (<$lfh>) {
  my ($name, $length) = split;

  # here we check that the sequence was in 'counts' file
  if (exists $sequence{$name}) { 
    $sequence{$name}{length} = $length;
  } 
}
close $lfh;

# and now the output block: it's mostly the same as in the original answer:
for my $name (sort keys %sequence) {
  print "$name, $sequence{$name}{count}, $sequence{$name}{length}", "\n";
}

Here's another codepad to illustrate how it works. Don't mind strange __DATA__ stuff, it's just for this specific version of program (using __DATA__ section allows me to simulate reading from files, as I can't use external sources at codepad).