Can float (or double) be set to NaN?

Question

Note: Similar to Can an integer be NaN in C++?

I understand this has little practical purpose, but can a float or double be set to NaN?

Gemmell

I wouldn't say there's little practical purpose. For instance I am using this method to specify failed results in a table of doubles. – Autoerotism

Answer 1

The Float object contains a static value, which is a float type, called NaN.

So

float myFloat = Float.NaN;

gives you what you are asking.

http://download.oracle.com/javase/6/docs/api/java/lang/Float.html#NaN

Breton

I thought that was just a constant representation for a wrapper class (but the constant is still a valid number)? I'm looking for setting the float or double to NaN, thus making it unusable in any arithmatic. – Gemmell

No, the constant is for the primitive type. – Breton

Ok, sorry, you edited before I finished this comment, thanks for the clarification. – Gemmell

Answer 2

Sure! NaN is a static constant in the Float and Double classes.

double x = Double.NaN;

Evenson

Answer 3

Yes

float f = Float.NaN;

See the doc for more info. Note that if you want to compare a number to NaN, you should use isNan().

Despite your question above, this does have a practical purpose. You can use this to indicate a value hasn't been set/provided yet.

Dovecote

Answer 4

jshell> double ddd=Double.NaN; ddd ==> NaN

jshell> ddd!=ddd $5 ==> true

Suffrage