I had this piece of code used somewhere but I hardly can understand it:

Breaking that down into stages:-

• Create an empty hash %seen that will track occurrences of elements in @output

• Pass each element of @output one at a time into the grep as $_ for filtering

• The first time a particular value occurs in $_ the hash value $seen{ $_ } will be empty, hence "false" and ! seen{ $_ } i.e. "not false" is "true" so that the $_ value passes out of the grep into @uniqueOutput

• Note also the ++ post-increment operator in ! $seen{ $_ }++ that increments the value of $seen{ $_ } after the test ! seen{ $_ } has been done, which means that after the first occurrence of a particular value the hash for it will no longer be blank, i.e. "false", but 1, 2 ... etc. depending on how many times it occurs, which evaluates to "true" and therefore "not true" is "false" so that second and subsequent occurrences will not pass from the grep into @uniqueOutput

My preference is to limit the scope of %seen to a do block so that it isn't left lying around.

johngg@abouriou:~/perl/Monks$ perl -Mstrict -Mwarnings -E '
say q{};
my @arr  = qw{ a b c d c e f a b b g d };
my @uniq = do {
   my %seen;
   grep { not $seen{ $_ } ++ } @arr;
   };
say qq{@uniq};'

a b c d e f g
[download]