I'm trying to better understand the Schwartzian Tranform (sorting a list by computable field). I've read the explanation in the Perl Cookbook, 2d ed. recipe 4.16. I also enjoyed Randal's 1996 article in the Unix Review, which describes a sortable-hash technique before it goes on to show the map-sort-map transform technique. The sortable-hash technique intrigues me. Other than coolness and nicer-looking code, what is the advantage of this:

    my @output =
      map { $_->[0] }
      sort { $a->[1] cmp $b->[1] }
      map { [$_, expensive_func($_)] }
    @input;
[download]

over this:

    foreach $_ (@input) {
        $result_for{$_} = expensive_func($_);
    }

    my @output
        = sort { $result_for{$a} cmp $result_for{$b}
               } @input;
[download]