Hi -

The use you state is valid.

They really come into their own when data needs to be able to identify when
it isn't holding a value. In SQL a statement could return a value or it
could return NULL. This doesn't cause too much problem for reference types
but obviously a value type needs a value (not unreasonable); this is where
the problem occurs.

You could assign an arbitrary NULL value (say 0) but it doesn't really model
NULL correctly as it's obviously in the range that the value holds.

Nullable types are created by the generic template struct Nullable and all
they essentially do is to maintain a boolean alongside the value indicating
whether it has been assigned to or not.

HTH

- Andy

That would seem to be a reasonable usage; more typical is representing data
from a database, where a null value has a specific meaning; otherwise you
need to rely on magic numbers such as "int.MinValue" or "-1" meaning "not
specified", or keep a separate "is <x> specified" flag. The nullable type
allows you to handle this directly, without risking messing up the database
by inserting a -1 by mistake.

Marc

Hi,

Basically that is the kind of situation you use them for.

Personally I think that the more common example is a date that is not set.
This is a fairly common escenario when dealiing with DBs