I feel like I need to write in defense of the "first approach." Not only is it "easier to follow, if you're new to Perl," but I think it's easier to follow even if you're not new to Perl.

Up to a point, I'd agree. Indeed, there's a reason I gave both approaches. Yet, you'll notice that it was when I went to the second approach that I felt the need to simplify the map (getting rid of the array indices -- this is Perl, not C), and then only after I'd done that did I really understand the logic of what was going on well enough to construct my fourth solution, which removes the nested map altogether. Having done that, we could combine it with the grep approach...

my %a = map { $_ => 1 } @a;
my @b = grep { $a{$_->[1]} } @d
[download]

This is perhaps clearest of all, if it does the right thing, which I suspect it does. I find it hard to believe that the original poster actually wanted the side-effect of matching substrings in this case.