On Saturday, July 6, 2013, Javier Guerra Giraldez wrote:
On Sat, Jul 6, 2013 at 12:07 PM, Jose Torre-Bueno <firstname.lastname@example.org> wrote:
> I am not understanding what is wrong with using nan as a place holder
> print (n1, n2, n1==n2, n1==n1)
-nan -nan false false
how do you use a placeholder that can't be equal to anything, not even
type(n1) ~= "string" and tostring(n1) == "-nan"
? Is it type number? If so, it's the only number not equal to itself, so you've got that going for it.
Interesting… NaN would actually work as a potential 'empty' then for my intended use, particularly as they can tunnel through the C layer also (though detecting them in a portable way is something I'd have to check up on carefully).