nan, nanf, nanl
From cppreference.com
                    
                                        
                    
                    
                                                            
                    |   Defined in header <math.h>
   | 
||
|   float       nanf( const char* arg );  | 
(since C99) | |
|   double      nan( const char* arg );  | 
(since C99) | |
|   long double nanl( const char* arg );  | 
(since C99) | |
Converts the implementation-defined character string arg into the corresponding quiet NaN value. The call std::nan("string") is equivalent to the call std::strtod("NAN(string)", (char**)NULL);.
[edit] Parameters
| arg | - | narrow character string identifying the contents of a NaN, or an empty string | 
[edit] Return value
The NaN value that corresponds to the identifying string arg or zero if the implemntation does not support quiet NaNs.
[edit] See also
|    (C99)  | 
   checks if the given number is NaN  (function)  | 
|   C++ documentation for nan 
 | |