Yer right, you want to use hashes:

# UNTESTED
# after creating @words and @all

for ( @all ) # iterate through @all, setting $_ to each in turn
{
  my $four = ( split("\t", $_) )[3]; # split each line on \t, extract 
+4th field
  $seen{$four}++; # add to hash
}

#  %seen hash has keys equal to the fourth field in 
# all the lines in @all

@matched = grep { exists $seen{$_} } @words;
@uniq = grep { not exists $seen{$_} } @words;

# the preceding 2 lines could also have been:

for ( @words )
{
   if ( exists $seen{$_} )
   {
      push @matched, $_;
   }
   else
   {
       push @uniq, $_;
   }
}
# which has the advantage of iterating over @words only once.
[download]

--Bob Niederman, http://bob-n.com