Legacy Document

floor

You can use the floor function to round a real number downward to the next integer value.

double_t floor (double_t x);

x

Any floating-point number.

DESCRIPTION

The floor function rounds its argument downward. This is an ANSI standard C library function. The result is returned in a floating-point data type.

This function is the same as performing the following code sequence:

r = fegetround();          /* save current rounding direction */
fesetround(FE_DOWNWARD);   /* round downward */
rint(x);                   /* round to integer */
fesetround(r);             /* restore rounding direction */

EXCEPTIONS

When x is finite and nonzero, the result of $floor(x)$ is exact.

SPECIAL CASES

Table 9-4 shows the results when the argument to the floor function is a zero, a NaN, or an Infinity.

Special cases for the floor function

Operation	Result	Exceptions raised
$floor(+0)$	+0	None
$floor(-0)$	$-0$	None
$floor(NaN)$	NaN	None[15]
$floor(+$ )	+	None
$floor(-$ )	$-$	None

EXAMPLES

z = floor(+INFINITY);   /* z = +INFINITY because +   is already an 
                           integer value by definition. */
z = floor(300.1);       /* z = 300.0 */
z = floor(-300.1);      /* z = -301.0 */

---

[15] If the NaN is a signaling NaN, the invalid exception is raised.