1:mod:`fractions` --- Rational numbers
2=====================================
3
4.. module:: fractions
5   :synopsis: Rational numbers.
6
7.. moduleauthor:: Jeffrey Yasskin <jyasskin at gmail.com>
8.. sectionauthor:: Jeffrey Yasskin <jyasskin at gmail.com>
9
10**Source code:** :source:`Lib/fractions.py`
11
12--------------
13
14The :mod:`fractions` module provides support for rational number arithmetic.
15
16
17A Fraction instance can be constructed from a pair of integers, from
18another rational number, or from a string.
19
20.. class:: Fraction(numerator=0, denominator=1)
21           Fraction(other_fraction)
22           Fraction(float)
23           Fraction(decimal)
24           Fraction(string)
25
26   The first version requires that *numerator* and *denominator* are instances
27   of :class:`numbers.Rational` and returns a new :class:`Fraction` instance
28   with value ``numerator/denominator``. If *denominator* is :const:`0`, it
29   raises a :exc:`ZeroDivisionError`. The second version requires that
30   *other_fraction* is an instance of :class:`numbers.Rational` and returns a
31   :class:`Fraction` instance with the same value.  The next two versions accept
32   either a :class:`float` or a :class:`decimal.Decimal` instance, and return a
33   :class:`Fraction` instance with exactly the same value.  Note that due to the
34   usual issues with binary floating-point (see :ref:`tut-fp-issues`), the
35   argument to ``Fraction(1.1)`` is not exactly equal to 11/10, and so
36   ``Fraction(1.1)`` does *not* return ``Fraction(11, 10)`` as one might expect.
37   (But see the documentation for the :meth:`limit_denominator` method below.)
38   The last version of the constructor expects a string or unicode instance.
39   The usual form for this instance is::
40
41      [sign] numerator ['/' denominator]
42
43   where the optional ``sign`` may be either '+' or '-' and
44   ``numerator`` and ``denominator`` (if present) are strings of
45   decimal digits.  In addition, any string that represents a finite
46   value and is accepted by the :class:`float` constructor is also
47   accepted by the :class:`Fraction` constructor.  In either form the
48   input string may also have leading and/or trailing whitespace.
49   Here are some examples::
50
51      >>> from fractions import Fraction
52      >>> Fraction(16, -10)
53      Fraction(-8, 5)
54      >>> Fraction(123)
55      Fraction(123, 1)
56      >>> Fraction()
57      Fraction(0, 1)
58      >>> Fraction('3/7')
59      Fraction(3, 7)
60      >>> Fraction(' -3/7 ')
61      Fraction(-3, 7)
62      >>> Fraction('1.414213 \t\n')
63      Fraction(1414213, 1000000)
64      >>> Fraction('-.125')
65      Fraction(-1, 8)
66      >>> Fraction('7e-6')
67      Fraction(7, 1000000)
68      >>> Fraction(2.25)
69      Fraction(9, 4)
70      >>> Fraction(1.1)
71      Fraction(2476979795053773, 2251799813685248)
72      >>> from decimal import Decimal
73      >>> Fraction(Decimal('1.1'))
74      Fraction(11, 10)
75
76
77   The :class:`Fraction` class inherits from the abstract base class
78   :class:`numbers.Rational`, and implements all of the methods and
79   operations from that class.  :class:`Fraction` instances are hashable,
80   and should be treated as immutable.  In addition,
81   :class:`Fraction` has the following properties and methods:
82
83   .. versionchanged:: 3.2
84      The :class:`Fraction` constructor now accepts :class:`float` and
85      :class:`decimal.Decimal` instances.
86
87   .. versionchanged:: 3.9
88      The :func:`math.gcd` function is now used to normalize the *numerator*
89      and *denominator*. :func:`math.gcd` always return a :class:`int` type.
90      Previously, the GCD type depended on *numerator* and *denominator*.
91
92   .. attribute:: numerator
93
94      Numerator of the Fraction in lowest term.
95
96   .. attribute:: denominator
97
98      Denominator of the Fraction in lowest term.
99
100
101   .. method:: as_integer_ratio()
102
103      Return a tuple of two integers, whose ratio is equal
104      to the Fraction and with a positive denominator.
105
106      .. versionadded:: 3.8
107
108   .. method:: from_float(flt)
109
110      This class method constructs a :class:`Fraction` representing the exact
111      value of *flt*, which must be a :class:`float`. Beware that
112      ``Fraction.from_float(0.3)`` is not the same value as ``Fraction(3, 10)``.
113
114      .. note::
115
116         From Python 3.2 onwards, you can also construct a
117         :class:`Fraction` instance directly from a :class:`float`.
118
119
120   .. method:: from_decimal(dec)
121
122      This class method constructs a :class:`Fraction` representing the exact
123      value of *dec*, which must be a :class:`decimal.Decimal` instance.
124
125      .. note::
126
127         From Python 3.2 onwards, you can also construct a
128         :class:`Fraction` instance directly from a :class:`decimal.Decimal`
129         instance.
130
131
132   .. method:: limit_denominator(max_denominator=1000000)
133
134      Finds and returns the closest :class:`Fraction` to ``self`` that has
135      denominator at most max_denominator.  This method is useful for finding
136      rational approximations to a given floating-point number:
137
138         >>> from fractions import Fraction
139         >>> Fraction('3.1415926535897932').limit_denominator(1000)
140         Fraction(355, 113)
141
142      or for recovering a rational number that's represented as a float:
143
144         >>> from math import pi, cos
145         >>> Fraction(cos(pi/3))
146         Fraction(4503599627370497, 9007199254740992)
147         >>> Fraction(cos(pi/3)).limit_denominator()
148         Fraction(1, 2)
149         >>> Fraction(1.1).limit_denominator()
150         Fraction(11, 10)
151
152
153   .. method:: __floor__()
154
155      Returns the greatest :class:`int` ``<= self``.  This method can
156      also be accessed through the :func:`math.floor` function:
157
158        >>> from math import floor
159        >>> floor(Fraction(355, 113))
160        3
161
162
163   .. method:: __ceil__()
164
165      Returns the least :class:`int` ``>= self``.  This method can
166      also be accessed through the :func:`math.ceil` function.
167
168
169   .. method:: __round__()
170               __round__(ndigits)
171
172      The first version returns the nearest :class:`int` to ``self``,
173      rounding half to even. The second version rounds ``self`` to the
174      nearest multiple of ``Fraction(1, 10**ndigits)`` (logically, if
175      ``ndigits`` is negative), again rounding half toward even.  This
176      method can also be accessed through the :func:`round` function.
177
178
179.. seealso::
180
181   Module :mod:`numbers`
182      The abstract base classes making up the numeric tower.
183