>> it can be implemented without touching HVM and we > > Ok. Sorry for OT :) > [ Remember that HB_IT_TIMESTAMP, HB_IT_DATETIME, could also be implemented > without touching HVM ]
Yes, most of them can be. But 'time' is a natural, general purpose data type. NULL on the contrary is a toy for SQL to make the theoretical/mathematical model happy and complete. Very debated if it's useful feature in SQL, and clearly not useful outside SQL. IOW it's not a generic, practical type. I'd also like to see how NIL can't replace NULL in context of SQL in practice.. I'd think that an RDD could simply return NIL instead of properly typed content, when the field content is NULL. Such behavior could also be controlled by DBI setting, plus we could have separate function to check whether a given field is NULL (regardless of DBI setting): HB_FIELDISNULL( <nPos> | <cName> ) -> <lNULL> Shortly I see we should definitely allow to access this information from .prg code, but it doesn't need a extra, special type. Brgds, Viktor _______________________________________________ Harbour mailing list (attachment size limit: 40KB) Harbour@harbour-project.org http://lists.harbour-project.org/mailman/listinfo/harbour
