I might be misunderstanding the question, but it sounds like you asking if you can assign $hash{'key'} = "something"; and then later assign  $hash{'key'} = "something_else";

This is not possible. You could put it in an array if it already exists, like $hash{'key'} = ( exists $hash{'key'} ) ? [ $hash{'key'}, "more" ] : "else"; but then you have to test each value before you use it. Unless, you just always put it in an array, ie $hash{'key'} = ( exists $hash{'key'} ) ? [ $hash{'key'}, "more" ] : ["else"];

Hope that wasn't too far off from the question...

--4c6966653a205468652073656172636820666f
  7220746861742070657266656374204765656b
  20436869632c20746865206f6e652077697468
  2074686520737061726b6c696e672065796573
  2077686f20706c617973206973207261696e2e

  if (defined $hash{$key}{$key}) {
    # do your thing
  }
[download]

My data is similar to:

File 1

dn=machinename
config=2000
speed=19200
setting=value1
setting=value2
setting=value3

File 2

dn=machinename2
config=2020
speed=9600
setting=value1
setting=value2
setting=value3


I need for all occurances of setting to end up in the hash for file one, and then compare the contents of file two and output the differences.

If the machines have different names (assuming a normal network :o)), don't you just need to do this? (continued from ZZamboni's code)

foreach $f (@files) {
  open FILE, $f or die "Error: $!";
  while(<FILE>) {
    chomp;
    my ($k,$v)=split("=",$_,2);
    # Assume dn= comes before other lines
    if ($k eq "dn") {
      push(@machines,$k);
      $machine = $k;
    } else {
       push(@{$machine}{$k},$v);
    }
  }
  close FILE;
}
[download]

In this case you would end up with almost the same effect:

@machines = ('machinename','machinename2');
${'machinename'}{'config'}   = 2000;
${'machinename'}{'speed'}    = 19600;
${'machinename'}{'setting'}  = value1, value2, value3;
${'machinename2'}{'config'}  = 2020;
${'machinename2'}{'speed'}   = 19600;
${'machinename2'}{'setting'} = value1, value2, value3;
etc...
[download]

Points being:
1. do you have to name a hash?
2. is it better to have @machines?

I'm starting to wonder about this too...

You could put all the values in array references indexed by your two-level hash. Something like this:

foreach $f (@files) {
  open FILE, $f or die "Error: $!";
  while(<FILE>) {
    chomp;
    my ($k,$v)=split("=",$_,2);
    # Assume dn= comes before other lines
    $level1=$k, next if $k eq "dn";
    unless ($level1) {
      die "Need to get dn= line first\n";
    }
    push @{$hash{$level1}{$k}}, $v;
  }
  close FILE;
}
[download]

Then you will have something like this:

  $hash{machinename}{config}=["2000"]
  $hash{machinename}{speed}=["19200"]
  $hash{machinename}{setting}=["value1", "value2", "value3"]
  $hash{machinename2}{config}=["2020"]
  etc.
[download]

And then do comparisons or anything else you need. See chapter 4 of The Perl Cookbook for examples of how to do fancy array operations, including comparisons.

I forgot to say, the value of dn will be the $level1 key and all of the other name=value pairs will be $level2 key = value.

Your humble servant,
-Chuck

So do you want a entry like:

dn=machinename1
name=value

to yield:

$hash{"machinename1"}{"name"}="value";

or

$hash{"machinename1"}{"name=value"}=1;

in case 1 you should save the different values for name in an array. something like this should work

foreach(@files){
    my $dn = "";
    open FILE, "$_" or die "error opening $_: ".$!;
    foreach(<FILE>){
	chomp;
        $dn = $1 if(/dn=(.*)/);
        push @{$hash{$dn}{$1}}, $2 if(/(^=*)=(.*)/ && $dn ne "");
    }
    close FILE;
}

if it is case2, I don't know why you would need to add the same line more than once. if you need the number of occurances try :

 
foreach(<FILE>){
    chomp;
    $dn = $1 if(/dn=(.*)/);
    $hash{$dn}{$_}++ if(/=)/ && $dn ne ""); 
}

for the inner loop. or use an array for the "name=value" entries:

foreach(<FILE>){
    chomp;
    $dn = $1 if(/dn=(.*)/);
    push @{$hash{$dn}}, $_ if(/=/ && $dn ne "");
}

hope that helps.