The Fractions
module allows us to create and work with rational numbers
: fractions with an integer numerator divided by an integer denominator.
For example, we can store 2/3
as an exact fraction instead of the approximate float
value 0.6666...
Unlike int
, float
, and complex
numbers, fractions do not have a literal form.
However, the fractions constructor is quite flexible.
Most obviously, it can take take two integers. Common factors are automatically removed, converting the fraction to its "lowest form": the smallest integers that accurately represent the fraction.
>>> from fractions import Fraction
>>> f1 = Fraction(2, 3) # 2/3
>>> f1
Fraction(2, 3)
>>> f2 = Fraction(6, 9)
>>> f2
Fraction(2, 3) # automatically simplified
>>> f1 == f2
True
The fractions constructor can also parse a string representation:
>>> f3 = Fraction('2/3')
>>> f3
Fraction(2, 3)
It can also work with float
parameters, but this may run into problems with the approximate nature of representing the decimal value internally as binary.
For more on this representation issue, see the 0.30000000000000004 website, and Floating Point Arithmetic: Issues and Limitations in the Python documentation.
For a more reliable result when using floats with fractions, there is the <fraction>.limit_denominator()
method.
.limit_denominator()
can take an integer parameter if you have specific requirements, but even the default (max_denominator=1000000
) can work well and give an acceptable, simple approximation.
>>> Fraction(1.2)
Fraction(5404319552844595, 4503599627370496)
>>> Fraction(1.2).limit_denominator()
Fraction(6, 5)
The usual arithmetic operators
+ - * / **
work with fractions, as with other numeric types.
Integers and other Fraction
s can be included and give a Fraction
result.
Including a float
in the expression results in float
output, with a consequent (possible) loss in precision.
>>> Fraction(2, 3) + Fraction(1, 4) # addition
Fraction(11, 12)
>>> Fraction(2, 3) * Fraction(6, 5) # multiply fractions
Fraction(4, 5)
>>> Fraction(2, 3) * 6 / 5 # fraction with integers
Fraction(4, 5)
>>> Fraction(2, 3) * 1.2 # fraction with float -> float
0.7999999999999999
>>> Fraction(2, 3) ** 2 # exponentiation with integer
Fraction(4, 9)
Fractions are great for preserving precision during intermediate calculations, but may not be what you want for the final output.
It is possible to get the numerator and denominator individually or as a tuple (tuples
will be discussed in a later Concept):
>>> Fraction(2, 3).numerator
2
>>> Fraction(2, 3).denominator
3
>>> Fraction(2, 3).as_integer_ratio()
(2, 3)
Various standard Python numeric functions also give the result you might expect from working with int
and float
types:
>>> round(Fraction(11, 3))
4
>>> from math import floor, ceil
>>> floor(Fraction(11, 3))
3
>>> ceil(Fraction(11, 3))
4
>>> float(Fraction(11, 3))
3.6666666666666665