> > Regarding: "So far the only suggestion was "use some non-deterministic > random sampling method and store the result", but since my samples are > large and I have lots of them, this would balloon my storage by >100x and I > don't have the available storage to make that work."
But Keith Medcalf was suggesting your table KeySet store only the *keys*, not the data itself, right? Are you saying storing the ROWIDs would still be prohibitively expensive? > _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users