Why float.NaN != double.NaN in C#?

Question

Why float.NaN != double.NaN ?

while float.PositiveInfinity == double.PositiveInfinity and float.NegativeInfinity == double.NegativeInfinity are equal.

EXAMPLE:

bool PosInfinity = (float.PositiveInfinity == double.PositiveInfinity); //true
bool NegInfinity = (float.NegativeInfinity == double.NegativeInfinity); //true

bool isNanEqual = (float.NaN == double.NaN);  //false, WHY?

Please, anyone give some examples when a number becomes NaN. — Javed Akram, Mar 06 '11 at 06:19
See also "Why is double.nan not equal to itself" http://stackoverflow.com/questions/1145443/why-is-double-nan-not-equal-to-itself — Philip Rieck, Mar 06 '11 at 16:54
Does this answer your question? [Why is double.NaN not equal to itself?](https://stackoverflow.com/questions/1145443/why-is-double-nan-not-equal-to-itself) — phuclv, Mar 02 '23 at 13:06

Damien_The_Unbeliever · Accepted Answer · 2011-03-06T06:55:04.257

NaN is never equal to NaN (even within the same type). Hence why the IsNaN function exists:

Double zero = 0;
// This will return true.
if (Double.IsNaN(0 / zero)) 
{
    Console.WriteLine("Double.IsNan() can determine whether a value is not-a-number.");
}

You should also be aware that none of the comparisons you've shown are actually occurring "as is". When you write floatValue == doubleValue, the floats will actually be implicitly converted to doubles before the comparison occurs.

score 29 · Answer 2 · answered Mar 06 '11 at 05:53

29

Probably because NaN != NaN

answered Mar 06 '11 at 05:53

Devin Jeanpierre

92,913
4
55
79

score 10 · Answer 3 · answered Mar 06 '11 at 05:53

10

To quote wikipedia:

A comparison with a NaN always returns an unordered result even when comparing with itself.

answered Mar 06 '11 at 05:53

Conrad Meyer

2,851
21
24

2

just like transact sql, NULL is allways different from NULL :P – aF. Nov 08 '11 at 13:46

Why float.NaN != double.NaN in C#?

3 Answers3