"OK, since you have got what sounds like mostly disjoint data (and do not care if you get some duplicates anyway) there is a workaround that I have used in Postgres which may be applicable:" . . . .