A two-pass process. First, count number of entries for each elem:

  my %item_count;
  %item_count{$_}++ foreach (@array);
[download]

Next, remove any element whose item_count != 1:

  @result = grep {$item_count{$_} == 1} @array;
[download]

--Dave.

Update: for the duplicate sub-pattern problem:

  my %item_count;
  %item_count{$_}++ for grep {/pattern/} @array
  @result = grep {$item_count{$_} <= 1} @array;
[download]

Update 2: to avoid the uninitialized value warning:

  @result = grep {($item_count{$_}||0) <= 1} @array;
[download]