Ox Predefined Constants

Constants defined in oxstd.oxh (requires #include <oxstd.oxh>)

FALSE           0
TRUE            1

WFL_DECFAILED   1
WFL_ITMAX       2
WFL_CONCAT      4

Constants defined in oxfloat.oxh (requires #include <oxfloat.oxh>)

M_PI            pi (3.1415...)
M_2PI           2*pi
M_PI_2          pi / 2
M_1_PI          1 / pi
M_SQRT2PI		sqrt(2*pi)
M_E             e = exp(1)
M_EULER         Euler's constant,
M_NAN           .NaN (Not a Number), 
                also see isnan() and isdotnan()
M_INF           .Inf, representing Infinity		
M_INF_POS       +.Inf, representing Infinity		
M_INF_NEG       -.Inf, representing minus Infinity		
DBL_DIG         number of decimal digits of precision
DBL_EPSILON     machine precision
DBL_MANT_DIG    number of bits in mantissa
DBL_MAX         maximum double value
DBL_MIN         minimum positive double value
DBL_MIN_EXP     minimum 2 exponent
DBL_MAX_EXP     maximum 2 exponent
DBL_MIN_E_EXP   minimum exp(.) exponent
DBL_MAX_E_EXP   maximum exp(.) exponent
DBL_MIN_10_EXP  minimum 10 exponent
DBL_MAX_10_EXP  maximum 10 exponent
INT_MAX         maximum integer value
INT_MIN         minimum integer value

The following constants are predefined by the Ox compiler:

OX_64_BIT      when running 64-bit Ox
OX_AIX         when running on IBM/AIX
OX_BIG_ENDIAN  only on a big-endian machine (Unix workstations)
OX_DecUNIX     when running on Dec/UNIX
OX_HPUX        when running on HP-UX
OX_Irix        when running on SGI/Irix
OX_Linux       when running on Linux/PC
OX_OS_X        when running on Mac/OS X
OX_PARALLEL    indicates that parallel and serial
               keywords are supported (Ox 7 onwards)
OX_Sun         when running on Sun
OX_Windows     when running under Windows

Missing values (NaN)

The hardware-defined missing value is called Not a Number, or .NaN for short. Any computation involving a .NaN results in a .NaN. The format used when printing output is .NaN.

In a matrix constant, either .NaN, M_NAN or a dot may be used to represent a missing value (M_NAN requires oxfloat.oxh). If the dot is the first or last element, an extra space is required to avoid confusion with dot-greater/less than.

In a double constant, either .NaN or M_NAN may be used to represent a missing value (M_NAN requires oxfloat.oxh). A number of procedures are available to deal with missing values, most importantly:

deletec(): deletes all columns which have a .NaN,
deleter(): deletes all rows which have a .NaN,
isdotnan(): returns matrix of 0's and 1's: 1 if the element is a .NaN, 0 otherwise,
isnan(): returns 1 if any element is a .NaN, 0 otherwise.
selectc(): selects all columns which have a .NaN,
selectr(): selects all rows which have a .NaN,

Infinity

Infinity also exists as a special value supported by the hardware. Infinity can be positive or negative (printed as +.Inf and -.Inf), and can be used in comparisons as any normal number. You can use .Inf, +.Inf and -.Inf in your code. Alternatively, the predefined constants M_INF, M_INF_POS, and M_INF_NEG are defined in oxfloat.oxh.


Ox version 7.00. © JA Doornik This file last changed .