MeowChow, here's another version using a hash slice instead of a for loop to assign the contents of one hash to another:

my %old = (
  1 => a,
  2 => b,
  3 => c,
  4 => d,
  5 => f,
);

my @keys = qw(4 5 6 7);  #Keys to copy from the old to the new hash

my %new;
@new{ @keys } = @old{ @keys };
[download]

In this case, the difference between using a for loop and a hash slice means a slight edge in speed during benchmarking. However, the larger the data set the wider the gap becomes.