the index to retrieve the tsvectors that contain at least N lexemes?
If not exploiting the index... any suggestion to improve performance
of such a query?
computing rank still requires retrieving a lot of tsvectors and
compute the rank for each of them.
Ivan Sergio Borgonovo