[ Arithmetic | Reference Manual | Alphabetic Index ]
# rational(+Number, -Result)

Converts Number into a rational number and unifies it with Result.
*Number*
- A number.
*Result*
- A variable or rational number.

## Description

This predicate is used by the ECLiPSe compiler to expand evaluable
arithmetic expressions. So the call to rational(Number, Result) is
equivalent to
Result is rational(Number)

which should be preferred.
When Number is an integer, Result is a rational with denominator 1.

When Number is already a rational, Result is identical to Number.

When Number is a float, Result is a rational whose value is exactly equal
to the value of the floating-point number. Since floats are usually
approximations of the intended value, the results may look unintuitive
and have unnecessarily large numerators and denominators. Use rationalize/2
to produce the most compact rational that still converts back into the
original float. rational/2 is usually faster than rationalize/2.

Bounded reals cannot be converted to rationals.

In coroutining mode, if Number is uninstantiated, the call to
rational/2 is delayed until this variable is instantiated.

### Modes and Determinism

### Exceptions

*(4) instantiation fault *
- Number is not instantiated (non-coroutining mode only).
*(24) number expected *
- Number is not of a numeric type.
*(141) unimplemented functionality *
- Number is a bounded real

## Examples

Success:
rational(25, 25_1).
rational(1.5, 3_2).
rational(3_4,3_4).
rational(9_12,3_4).
rational(-6, Result). (gives Result = -6_1)
rational(0.1, Result). (gives Result = 3602879701896397_36028797018963968)
Fail:
rational(1, 2_1).
rational(3, 3).
rational(1, r).
Error:
rational(A, 1_3). (Error 4).
rational(4 + 2, 6_1). (Error 24).
rational(0.9__1.1, X). (Error 141).

## See Also

rationalize / 2, is / 2