Numerical Type Conversions

In what follows we will use the short notation I,R,D,C for INTEGER, REAL, DOUBLE PRECISION and COMPLEX.

• INT(x); converts a R, D or C argument (x) to the corresponding INTEGER approximation using ``round towards '' (dumps the fractional part). For a complex argument, INT(real part) is taken.
• CEILING(x); converts a R or D argument (x) to the corresponding INTEGER approximation using ``round up''.
• FLOOR(x); converts a R or D argument (x) to the corresponding INTEGER approximation using ``round down''.
• NINT(x); converts a R or D argument (x) to the nearest INTEGER (using ``round to nearest'').

For example,

• REAL(k); converts a I or D argument (k) to the corresponding REAL approximation. Selects the real part of a C argument.
• AIMAG(z); selects the imaginary part of a C argument.
• DBLE(x); converts an INTEGER or REAL argument (x) to the corresponding DOUBLE PRECISION approximation.
• CMPLX(x,y); returns the complex value x + iy. The arguments x,y are usually REAL, but can also be INTEGER or DOUBLE PRECISION, in which case the returned value is REAL(x) + iREAL(y). CMPLX(x) returns the complex value REAL(x) + i.

For example,

{}
integer :: i=2
real ::    t=1.7
double precision :: x=1.7d0,y=2.8d0
complex :: z=(1.7,2.8)
print*, x, real(x) ! output = 1.7,  1.70000005
print*, real(z), aimag(z) ! output = 1.70000005,  2.79999995
print*, dble(t) ! output = 1.7000000476837158
print*, cmplx(i), cmplx(z) ! output = (2.,0.e+0),  (1.70000005,2.79999995)

Note the errors in double to real and real to double conversions, as well as the representation errors.