That's a very cool way of solving this problem, I learned quite a bit from reading it. I noticed that this returns lines that contain at least one element in @arr. I made a simple modification to the second algorithm to only return lines that have all elements in @arr. The OP didn't specify if he meant one or all array elements, so here's my modification for brevity's sake.

my (%hsh, @keepers);
@hsh{@arr} = ();

while (<>) {
    push @keepers, $_ unless grep {!exists $hsh{$_}} split;
}
[download]

Thanks for enlightening me with that code sample Zaxo