I did it myself before peeking at the above answers, and came up with something that avoids grep and seems pretty concise to me, here it is:

use strict;
use warnings;

my @big = qw(a b c d e f g h i);
my @small = qw(f c d g);
my %small; $small{$_}=1 for @small;

my @keys_wanted = @small;
foreach ( @big ) {
    push @keys_wanted, $_ unless $small{$_};
}

print "@keys_wanted";
[download]

This doesn't remove dupes though (see ikegami's comments above), and I'm not sure if it can be simply modified to do so.