While your algorithm works (mostly), it trades a small memory gain for a lot of extra CPU work, especially as @do_not_want grows.

For every key in %hash, the code must loop through every member of @do_not_want. Grep does not short-circuit, so even after you've found 'ytd', you are still going to check 'mtd' and 'wtd'.

The other problem with this code is that if $hash{ytd_total} = 1000; is a desired report value, your grep will match and incorrectly exclude that value.

As long as pure textual equivalence is the matching criteria, using a hash for the lookup is going to be faster and make a tiny impact on memory.

If you need to do some kind of complex matching, then there is no way (that I know of) to avoid nested loops. But you can use a short circuited loop to avoid checking for any matches after the first. List::MoreUtils any comes to mind as an excellent way to handle this situation.

Here's a version that uses a hash

my %hash = (
        ytd => 1.5,
        mtd => 2.0,
        wtd => 2.5,
        Jumbo_Tron => "United Center",
        Meat_Pie => "Gross",
        "Word" => "Association",
        );
my %do_not_want;
@do_not_want{ qw(ytd mtd wtd) } = ();

foreach my $k (keys %hash) {
        next if exists $do_not_want{$k};
        print "$k => $hash{$k}\n";
}
[download]

Here's a version that matches hash keys against a regex:

use List::MoreUtils qw(any);

my %hash = (
        ytd => 1.5,
        mtd => 2.0,
        wtd => 2.5,
        Jumbo_Tron => "United Center",
        Meat_Pie => "Gross",
        "Word" => "Association",
        );
my @do_not_want = ( qr/ytd/, qr/mtd/, qr/wtd/ );

foreach my $k (keys %hash) {
        next if any { $k =~ $_ } @do_not_want ;
        print "$k => $hash{$k}\n";
}
[download]

Warning, all this code is untested. I modified it in a browser edit box.