2 :mod:`collections` --- High-performance container datatypes
3 ===========================================================
5 .. module:: collections
6 :synopsis: High-performance datatypes
7 .. moduleauthor:: Raymond Hettinger <python@rcn.com>
8 .. sectionauthor:: Raymond Hettinger <python@rcn.com>
14 from collections import *
16 __name__ = '<doctest>'
18 This module implements high-performance container datatypes. Currently,
19 there are four datatypes, :class:`Counter`, :class:`deque`, :class:`OrderedDict` and
20 :class:`defaultdict`, and one datatype factory function, :func:`namedtuple`.
22 The specialized containers provided in this module provide alternatives
23 to Python's general purpose built-in containers, :class:`dict`,
24 :class:`list`, :class:`set`, and :class:`tuple`.
26 .. versionchanged:: 2.4
29 .. versionchanged:: 2.5
30 Added :class:`defaultdict`.
32 .. versionchanged:: 2.6
33 Added :func:`namedtuple` and added abstract base classes.
35 .. versionchanged:: 2.7
36 Added :class:`Counter` and :class:`OrderedDict`.
38 In addition to containers, the collections module provides some ABCs
39 (abstract base classes) that can be used to test whether a class
40 provides a particular interface, for example, whether it is hashable or
44 ABCs - abstract base classes
45 ----------------------------
47 The collections module offers the following ABCs:
49 ========================= ===================== ====================== ====================================================
50 ABC Inherits Abstract Methods Mixin Methods
51 ========================= ===================== ====================== ====================================================
52 :class:`Container` ``__contains__``
53 :class:`Hashable` ``__hash__``
54 :class:`Iterable` ``__iter__``
55 :class:`Iterator` :class:`Iterable` ``__next__`` ``__iter__``
56 :class:`Sized` ``__len__``
57 :class:`Callable` ``__call__``
59 :class:`Sequence` :class:`Sized`, ``__getitem__`` ``__contains__``. ``__iter__``, ``__reversed__``.
60 :class:`Iterable`, ``index``, and ``count``
63 :class:`MutableSequence` :class:`Sequence` ``__setitem__`` Inherited Sequence methods and
64 ``__delitem__``, ``append``, ``reverse``, ``extend``, ``pop``,
65 and ``insert`` ``remove``, and ``__iadd__``
67 :class:`Set` :class:`Sized`, ``__le__``, ``__lt__``, ``__eq__``, ``__ne__``,
68 :class:`Iterable`, ``__gt__``, ``__ge__``, ``__and__``, ``__or__``
69 :class:`Container` ``__sub__``, ``__xor__``, and ``isdisjoint``
71 :class:`MutableSet` :class:`Set` ``add`` and Inherited Set methods and
72 ``discard`` ``clear``, ``pop``, ``remove``, ``__ior__``,
73 ``__iand__``, ``__ixor__``, and ``__isub__``
75 :class:`Mapping` :class:`Sized`, ``__getitem__`` ``__contains__``, ``keys``, ``items``, ``values``,
76 :class:`Iterable`, ``get``, ``__eq__``, and ``__ne__``
79 :class:`MutableMapping` :class:`Mapping` ``__setitem__`` and Inherited Mapping methods and
80 ``__delitem__`` ``pop``, ``popitem``, ``clear``, ``update``,
84 :class:`MappingView` :class:`Sized` ``__len__``
85 :class:`KeysView` :class:`MappingView`, ``__contains__``,
86 :class:`Set` ``__iter__``
87 :class:`ItemsView` :class:`MappingView`, ``__contains__``,
88 :class:`Set` ``__iter__``
89 :class:`ValuesView` :class:`MappingView` ``__contains__``, ``__iter__``
90 ========================= ===================== ====================== ====================================================
92 These ABCs allow us to ask classes or instances if they provide
93 particular functionality, for example::
96 if isinstance(myvar, collections.Sized):
99 Several of the ABCs are also useful as mixins that make it easier to develop
100 classes supporting container APIs. For example, to write a class supporting
101 the full :class:`Set` API, it only necessary to supply the three underlying
102 abstract methods: :meth:`__contains__`, :meth:`__iter__`, and :meth:`__len__`.
103 The ABC supplies the remaining methods such as :meth:`__and__` and
104 :meth:`isdisjoint` ::
106 class ListBasedSet(collections.Set):
107 ''' Alternate set implementation favoring space over speed
108 and not requiring the set elements to be hashable. '''
109 def __init__(self, iterable):
110 self.elements = lst = []
111 for value in iterable:
115 return iter(self.elements)
116 def __contains__(self, value):
117 return value in self.elements
119 return len(self.elements)
121 s1 = ListBasedSet('abcdef')
122 s2 = ListBasedSet('defghi')
123 overlap = s1 & s2 # The __and__() method is supported automatically
125 Notes on using :class:`Set` and :class:`MutableSet` as a mixin:
128 Since some set operations create new sets, the default mixin methods need
129 a way to create new instances from an iterable. The class constructor is
130 assumed to have a signature in the form ``ClassName(iterable)``.
131 That assumption is factored-out to an internal classmethod called
132 :meth:`_from_iterable` which calls ``cls(iterable)`` to produce a new set.
133 If the :class:`Set` mixin is being used in a class with a different
134 constructor signature, you will need to override :meth:`from_iterable`
135 with a classmethod that can construct new instances from
136 an iterable argument.
139 To override the comparisons (presumably for speed, as the
140 semantics are fixed), redefine :meth:`__le__` and
141 then the other operations will automatically follow suit.
144 The :class:`Set` mixin provides a :meth:`_hash` method to compute a hash value
145 for the set; however, :meth:`__hash__` is not defined because not all sets
146 are hashable or immutable. To add set hashabilty using mixins,
147 inherit from both :meth:`Set` and :meth:`Hashable`, then define
148 ``__hash__ = Set._hash``.
152 * `OrderedSet recipe <http://code.activestate.com/recipes/576694/>`_ for an
153 example built on :class:`MutableSet`.
155 * For more about ABCs, see the :mod:`abc` module and :pep:`3119`.
158 :class:`Counter` objects
159 ------------------------
161 A counter tool is provided to support convenient and rapid tallies.
164 >>> # Tally occurrences of words in a list
166 >>> for word in ['red', 'blue', 'red', 'green', 'blue', 'blue']:
169 Counter({'blue': 3, 'red': 2, 'green': 1})
171 >>> # Find the ten most common words in Hamlet
173 >>> words = re.findall('\w+', open('hamlet.txt').read().lower())
174 >>> Counter(words).most_common(10)
175 [('the', 1143), ('and', 966), ('to', 762), ('of', 669), ('i', 631),
176 ('you', 554), ('a', 546), ('my', 514), ('hamlet', 471), ('in', 451)]
178 .. class:: Counter([iterable-or-mapping])
180 A :class:`Counter` is a :class:`dict` subclass for counting hashable objects.
181 It is an unordered collection where elements are stored as dictionary keys
182 and their counts are stored as dictionary values. Counts are allowed to be
183 any integer value including zero or negative counts. The :class:`Counter`
184 class is similar to bags or multisets in other languages.
186 Elements are counted from an *iterable* or initialized from another
187 *mapping* (or counter)::
189 >>> c = Counter() # a new, empty counter
190 >>> c = Counter('gallahad') # a new counter from an iterable
191 >>> c = Counter({'red': 4, 'blue': 2}) # a new counter from a mapping
192 >>> c = Counter(cats=4, dogs=8) # a new counter from keyword args
194 Counter objects have a dictionary interface except that they return a zero
195 count for missing items instead of raising a :exc:`KeyError`::
197 >>> c = Counter(['eggs', 'ham'])
198 >>> c['bacon'] # count of a missing element is zero
201 Setting a count to zero does not remove an element from a counter.
202 Use ``del`` to remove it entirely:
204 >>> c['sausage'] = 0 # counter entry with a zero count
205 >>> del c['sausage'] # del actually removes the entry
207 .. versionadded:: 2.7
210 Counter objects support two methods beyond those available for all
213 .. method:: elements()
215 Return an iterator over elements repeating each as many times as its
216 count. Elements are returned in arbitrary order. If an element's count
217 is less than one, :meth:`elements` will ignore it.
219 >>> c = Counter(a=4, b=2, c=0, d=-2)
220 >>> list(c.elements())
221 ['a', 'a', 'a', 'a', 'b', 'b']
223 .. method:: most_common([n])
225 Return a list of the *n* most common elements and their counts from the
226 most common to the least. If *n* is not specified, :func:`most_common`
227 returns *all* elements in the counter. Elements with equal counts are
228 ordered arbitrarily::
230 >>> Counter('abracadabra').most_common(3)
231 [('a', 5), ('r', 2), ('b', 2)]
233 The usual dictionary methods are available for :class:`Counter` objects
234 except for two which work differently for counters.
236 .. method:: fromkeys(iterable)
238 This class method is not implemented for :class:`Counter` objects.
240 .. method:: update([iterable-or-mapping])
242 Elements are counted from an *iterable* or added-in from another
243 *mapping* (or counter). Like :meth:`dict.update` but adds counts
244 instead of replacing them. Also, the *iterable* is expected to be a
245 sequence of elements, not a sequence of ``(key, value)`` pairs.
247 Common patterns for working with :class:`Counter` objects::
249 sum(c.values()) # total of all counts
250 c.clear() # reset all counts
251 list(c) # list unique elements
252 set(c) # convert to a set
253 dict(c) # convert to a regular dictionary
254 c.items() # convert to a list of (elem, cnt) pairs
255 Counter(dict(list_of_pairs)) # convert from a list of (elem, cnt) pairs
256 c.most_common()[:-n:-1] # n least common elements
257 c += Counter() # remove zero and negative counts
259 Several mathematical operations are provided for combining :class:`Counter`
260 objects to produce multisets (counters that have counts greater than zero).
261 Addition and subtraction combine counters by adding or subtracting the counts
262 of corresponding elements. Intersection and union return the minimum and
263 maximum of corresponding counts. Each operation can accept inputs with signed
264 counts, but the output will exclude results with counts of zero or less.
266 >>> c = Counter(a=3, b=1)
267 >>> d = Counter(a=1, b=2)
268 >>> c + d # add two counters together: c[x] + d[x]
269 Counter({'a': 4, 'b': 3})
270 >>> c - d # subtract (keeping only positive counts)
272 >>> c & d # intersection: min(c[x], d[x])
273 Counter({'a': 1, 'b': 1})
274 >>> c | d # union: max(c[x], d[x])
275 Counter({'a': 3, 'b': 2})
279 * `Counter class <http://code.activestate.com/recipes/576611/>`_
280 adapted for Python 2.5 and an early `Bag recipe
281 <http://code.activestate.com/recipes/259174/>`_ for Python 2.4.
283 * `Bag class <http://www.gnu.org/software/smalltalk/manual-base/html_node/Bag.html>`_
286 * Wikipedia entry for `Multisets <http://en.wikipedia.org/wiki/Multiset>`_\.
288 * `C++ multisets <http://www.demo2s.com/Tutorial/Cpp/0380__set-multiset/Catalog0380__set-multiset.htm>`_
289 tutorial with examples.
291 * For mathematical operations on multisets and their use cases, see
292 *Knuth, Donald. The Art of Computer Programming Volume II,
293 Section 4.6.3, Exercise 19*\.
295 * To enumerate all distinct multisets of a given size over a given set of
296 elements, see :func:`itertools.combinations_with_replacement`.
298 map(Counter, combinations_with_replacement('ABC', 2)) --> AA AB AC BB BC CC
301 :class:`deque` objects
302 ----------------------
304 .. class:: deque([iterable[, maxlen]])
306 Returns a new deque object initialized left-to-right (using :meth:`append`) with
307 data from *iterable*. If *iterable* is not specified, the new deque is empty.
309 Deques are a generalization of stacks and queues (the name is pronounced "deck"
310 and is short for "double-ended queue"). Deques support thread-safe, memory
311 efficient appends and pops from either side of the deque with approximately the
312 same O(1) performance in either direction.
314 Though :class:`list` objects support similar operations, they are optimized for
315 fast fixed-length operations and incur O(n) memory movement costs for
316 ``pop(0)`` and ``insert(0, v)`` operations which change both the size and
317 position of the underlying data representation.
319 .. versionadded:: 2.4
321 If *maxlen* is not specified or is *None*, deques may grow to an
322 arbitrary length. Otherwise, the deque is bounded to the specified maximum
323 length. Once a bounded length deque is full, when new items are added, a
324 corresponding number of items are discarded from the opposite end. Bounded
325 length deques provide functionality similar to the ``tail`` filter in
326 Unix. They are also useful for tracking transactions and other pools of data
327 where only the most recent activity is of interest.
329 .. versionchanged:: 2.6
330 Added *maxlen* parameter.
332 Deque objects support the following methods:
335 .. method:: append(x)
337 Add *x* to the right side of the deque.
340 .. method:: appendleft(x)
342 Add *x* to the left side of the deque.
347 Remove all elements from the deque leaving it with length 0.
350 .. method:: extend(iterable)
352 Extend the right side of the deque by appending elements from the iterable
356 .. method:: extendleft(iterable)
358 Extend the left side of the deque by appending elements from *iterable*.
359 Note, the series of left appends results in reversing the order of
360 elements in the iterable argument.
365 Remove and return an element from the right side of the deque. If no
366 elements are present, raises an :exc:`IndexError`.
369 .. method:: popleft()
371 Remove and return an element from the left side of the deque. If no
372 elements are present, raises an :exc:`IndexError`.
375 .. method:: remove(value)
377 Removed the first occurrence of *value*. If not found, raises a
380 .. versionadded:: 2.5
383 .. method:: rotate(n)
385 Rotate the deque *n* steps to the right. If *n* is negative, rotate to
386 the left. Rotating one step to the right is equivalent to:
387 ``d.appendleft(d.pop())``.
390 Deque objects also provide one read-only attribute:
392 .. attribute:: maxlen
394 Maximum size of a deque or *None* if unbounded.
396 .. versionadded:: 2.7
399 In addition to the above, deques support iteration, pickling, ``len(d)``,
400 ``reversed(d)``, ``copy.copy(d)``, ``copy.deepcopy(d)``, membership testing with
401 the :keyword:`in` operator, and subscript references such as ``d[-1]``. Indexed
402 access is O(1) at both ends but slows to O(n) in the middle. For fast random
403 access, use lists instead.
409 >>> from collections import deque
410 >>> d = deque('ghi') # make a new deque with three items
411 >>> for elem in d: # iterate over the deque's elements
412 ... print elem.upper()
417 >>> d.append('j') # add a new entry to the right side
418 >>> d.appendleft('f') # add a new entry to the left side
419 >>> d # show the representation of the deque
420 deque(['f', 'g', 'h', 'i', 'j'])
422 >>> d.pop() # return and remove the rightmost item
424 >>> d.popleft() # return and remove the leftmost item
426 >>> list(d) # list the contents of the deque
428 >>> d[0] # peek at leftmost item
430 >>> d[-1] # peek at rightmost item
433 >>> list(reversed(d)) # list the contents of a deque in reverse
435 >>> 'h' in d # search the deque
437 >>> d.extend('jkl') # add multiple elements at once
439 deque(['g', 'h', 'i', 'j', 'k', 'l'])
440 >>> d.rotate(1) # right rotation
442 deque(['l', 'g', 'h', 'i', 'j', 'k'])
443 >>> d.rotate(-1) # left rotation
445 deque(['g', 'h', 'i', 'j', 'k', 'l'])
447 >>> deque(reversed(d)) # make a new deque in reverse order
448 deque(['l', 'k', 'j', 'i', 'h', 'g'])
449 >>> d.clear() # empty the deque
450 >>> d.pop() # cannot pop from an empty deque
451 Traceback (most recent call last):
452 File "<pyshell#6>", line 1, in -toplevel-
454 IndexError: pop from an empty deque
456 >>> d.extendleft('abc') # extendleft() reverses the input order
458 deque(['c', 'b', 'a'])
461 :class:`deque` Recipes
462 ^^^^^^^^^^^^^^^^^^^^^^
464 This section shows various approaches to working with deques.
466 Bounded length deques provide functionality similar to the ``tail`` filter
469 def tail(filename, n=10):
470 'Return the last n lines of a file'
471 return deque(open(filename), n)
473 Another approach to using deques is to maintain a sequence of recently
474 added elements by appending to the right and popping to the left::
476 def moving_average(iterable, n=3):
477 # moving_average([40, 30, 50, 46, 39, 44]) --> 40.0 42.0 45.0 43.0
478 # http://en.wikipedia.org/wiki/Moving_average
481 d = deque(itertools.islice(it, n))
486 s += elem - d.popleft()
490 The :meth:`rotate` method provides a way to implement :class:`deque` slicing and
491 deletion. For example, a pure python implementation of ``del d[n]`` relies on
492 the :meth:`rotate` method to position elements to be popped::
494 def delete_nth(d, n):
499 To implement :class:`deque` slicing, use a similar approach applying
500 :meth:`rotate` to bring a target element to the left side of the deque. Remove
501 old entries with :meth:`popleft`, add new entries with :meth:`extend`, and then
502 reverse the rotation.
503 With minor variations on that approach, it is easy to implement Forth style
504 stack manipulations such as ``dup``, ``drop``, ``swap``, ``over``, ``pick``,
505 ``rot``, and ``roll``.
508 :class:`defaultdict` objects
509 ----------------------------
511 .. class:: defaultdict([default_factory[, ...]])
513 Returns a new dictionary-like object. :class:`defaultdict` is a subclass of the
514 builtin :class:`dict` class. It overrides one method and adds one writable
515 instance variable. The remaining functionality is the same as for the
516 :class:`dict` class and is not documented here.
518 The first argument provides the initial value for the :attr:`default_factory`
519 attribute; it defaults to ``None``. All remaining arguments are treated the same
520 as if they were passed to the :class:`dict` constructor, including keyword
523 .. versionadded:: 2.5
525 :class:`defaultdict` objects support the following method in addition to the
526 standard :class:`dict` operations:
529 .. method:: defaultdict.__missing__(key)
531 If the :attr:`default_factory` attribute is ``None``, this raises a
532 :exc:`KeyError` exception with the *key* as argument.
534 If :attr:`default_factory` is not ``None``, it is called without arguments
535 to provide a default value for the given *key*, this value is inserted in
536 the dictionary for the *key*, and returned.
538 If calling :attr:`default_factory` raises an exception this exception is
539 propagated unchanged.
541 This method is called by the :meth:`__getitem__` method of the
542 :class:`dict` class when the requested key is not found; whatever it
543 returns or raises is then returned or raised by :meth:`__getitem__`.
546 :class:`defaultdict` objects support the following instance variable:
549 .. attribute:: defaultdict.default_factory
551 This attribute is used by the :meth:`__missing__` method; it is
552 initialized from the first argument to the constructor, if present, or to
556 :class:`defaultdict` Examples
557 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
559 Using :class:`list` as the :attr:`default_factory`, it is easy to group a
560 sequence of key-value pairs into a dictionary of lists:
562 >>> s = [('yellow', 1), ('blue', 2), ('yellow', 3), ('blue', 4), ('red', 1)]
563 >>> d = defaultdict(list)
568 [('blue', [2, 4]), ('red', [1]), ('yellow', [1, 3])]
570 When each key is encountered for the first time, it is not already in the
571 mapping; so an entry is automatically created using the :attr:`default_factory`
572 function which returns an empty :class:`list`. The :meth:`list.append`
573 operation then attaches the value to the new list. When keys are encountered
574 again, the look-up proceeds normally (returning the list for that key) and the
575 :meth:`list.append` operation adds another value to the list. This technique is
576 simpler and faster than an equivalent technique using :meth:`dict.setdefault`:
580 ... d.setdefault(k, []).append(v)
583 [('blue', [2, 4]), ('red', [1]), ('yellow', [1, 3])]
585 Setting the :attr:`default_factory` to :class:`int` makes the
586 :class:`defaultdict` useful for counting (like a bag or multiset in other
589 >>> s = 'mississippi'
590 >>> d = defaultdict(int)
595 [('i', 4), ('p', 2), ('s', 4), ('m', 1)]
597 When a letter is first encountered, it is missing from the mapping, so the
598 :attr:`default_factory` function calls :func:`int` to supply a default count of
599 zero. The increment operation then builds up the count for each letter.
601 The function :func:`int` which always returns zero is just a special case of
602 constant functions. A faster and more flexible way to create constant functions
603 is to use :func:`itertools.repeat` which can supply any constant value (not just
606 >>> def constant_factory(value):
607 ... return itertools.repeat(value).next
608 >>> d = defaultdict(constant_factory('<missing>'))
609 >>> d.update(name='John', action='ran')
610 >>> '%(name)s %(action)s to %(object)s' % d
611 'John ran to <missing>'
613 Setting the :attr:`default_factory` to :class:`set` makes the
614 :class:`defaultdict` useful for building a dictionary of sets:
616 >>> s = [('red', 1), ('blue', 2), ('red', 3), ('blue', 4), ('red', 1), ('blue', 4)]
617 >>> d = defaultdict(set)
622 [('blue', set([2, 4])), ('red', set([1, 3]))]
625 :func:`namedtuple` Factory Function for Tuples with Named Fields
626 ----------------------------------------------------------------
628 Named tuples assign meaning to each position in a tuple and allow for more readable,
629 self-documenting code. They can be used wherever regular tuples are used, and
630 they add the ability to access fields by name instead of position index.
632 .. function:: namedtuple(typename, field_names, [verbose], [rename])
634 Returns a new tuple subclass named *typename*. The new subclass is used to
635 create tuple-like objects that have fields accessible by attribute lookup as
636 well as being indexable and iterable. Instances of the subclass also have a
637 helpful docstring (with typename and field_names) and a helpful :meth:`__repr__`
638 method which lists the tuple contents in a ``name=value`` format.
640 The *field_names* are a single string with each fieldname separated by whitespace
641 and/or commas, for example ``'x y'`` or ``'x, y'``. Alternatively, *field_names*
642 can be a sequence of strings such as ``['x', 'y']``.
644 Any valid Python identifier may be used for a fieldname except for names
645 starting with an underscore. Valid identifiers consist of letters, digits,
646 and underscores but do not start with a digit or underscore and cannot be
647 a :mod:`keyword` such as *class*, *for*, *return*, *global*, *pass*, *print*,
650 If *rename* is true, invalid fieldnames are automatically replaced
651 with positional names. For example, ``['abc', 'def', 'ghi', 'abc']`` is
652 converted to ``['abc', '_1', 'ghi', '_3']``, eliminating the keyword
653 ``def`` and the duplicate fieldname ``abc``.
655 If *verbose* is true, the class definition is printed just before being built.
657 Named tuple instances do not have per-instance dictionaries, so they are
658 lightweight and require no more memory than regular tuples.
660 .. versionadded:: 2.6
662 .. versionchanged:: 2.7
663 added support for *rename*.
668 :options: +NORMALIZE_WHITESPACE
670 >>> Point = namedtuple('Point', 'x y', verbose=True)
678 def __new__(cls, x, y):
679 return tuple.__new__(cls, (x, y))
682 def _make(cls, iterable, new=tuple.__new__, len=len):
683 'Make a new Point object from a sequence or iterable'
684 result = new(cls, iterable)
686 raise TypeError('Expected 2 arguments, got %d' % len(result))
690 return 'Point(x=%r, y=%r)' % self
693 'Return a new OrderedDict which maps field names to their values'
694 return OrderedDict(zip(self._fields, self))
696 def _replace(self, **kwds):
697 'Return a new Point object replacing specified fields with new values'
698 result = self._make(map(kwds.pop, ('x', 'y'), self))
700 raise ValueError('Got unexpected field names: %r' % kwds.keys())
703 def __getnewargs__(self):
706 x = property(itemgetter(0))
707 y = property(itemgetter(1))
709 >>> p = Point(11, y=22) # instantiate with positional or keyword arguments
710 >>> p[0] + p[1] # indexable like the plain tuple (11, 22)
712 >>> x, y = p # unpack like a regular tuple
715 >>> p.x + p.y # fields also accessible by name
717 >>> p # readable __repr__ with a name=value style
720 Named tuples are especially useful for assigning field names to result tuples returned
721 by the :mod:`csv` or :mod:`sqlite3` modules::
723 EmployeeRecord = namedtuple('EmployeeRecord', 'name, age, title, department, paygrade')
726 for emp in map(EmployeeRecord._make, csv.reader(open("employees.csv", "rb"))):
727 print emp.name, emp.title
730 conn = sqlite3.connect('/companydata')
731 cursor = conn.cursor()
732 cursor.execute('SELECT name, age, title, department, paygrade FROM employees')
733 for emp in map(EmployeeRecord._make, cursor.fetchall()):
734 print emp.name, emp.title
736 In addition to the methods inherited from tuples, named tuples support
737 three additional methods and one attribute. To prevent conflicts with
738 field names, the method and attribute names start with an underscore.
740 .. method:: somenamedtuple._make(iterable)
742 Class method that makes a new instance from an existing sequence or iterable.
750 .. method:: somenamedtuple._asdict()
752 Return a new :class:`OrderedDict` which maps field names to their corresponding
756 OrderedDict([('x', 11), ('y', 22)])
758 .. versionchanged:: 2.7
759 Returns an :class:`OrderedDict` instead of a regular :class:`dict`.
761 .. method:: somenamedtuple._replace(kwargs)
763 Return a new instance of the named tuple replacing specified fields with new
766 >>> p = Point(x=11, y=22)
770 >>> for partnum, record in inventory.items():
771 ... inventory[partnum] = record._replace(price=newprices[partnum], timestamp=time.now())
773 .. attribute:: somenamedtuple._fields
775 Tuple of strings listing the field names. Useful for introspection
776 and for creating new named tuple types from existing named tuples.
780 >>> p._fields # view the field names
783 >>> Color = namedtuple('Color', 'red green blue')
784 >>> Pixel = namedtuple('Pixel', Point._fields + Color._fields)
785 >>> Pixel(11, 22, 128, 255, 0)
786 Pixel(x=11, y=22, red=128, green=255, blue=0)
788 To retrieve a field whose name is stored in a string, use the :func:`getattr`
794 To convert a dictionary to a named tuple, use the double-star-operator
795 (as described in :ref:`tut-unpacking-arguments`):
797 >>> d = {'x': 11, 'y': 22}
801 Since a named tuple is a regular Python class, it is easy to add or change
802 functionality with a subclass. Here is how to add a calculated field and
803 a fixed-width print format:
805 >>> class Point(namedtuple('Point', 'x y')):
809 ... return (self.x ** 2 + self.y ** 2) ** 0.5
810 ... def __str__(self):
811 ... return 'Point: x=%6.3f y=%6.3f hypot=%6.3f' % (self.x, self.y, self.hypot)
813 >>> for p in Point(3, 4), Point(14, 5/7.):
815 Point: x= 3.000 y= 4.000 hypot= 5.000
816 Point: x=14.000 y= 0.714 hypot=14.018
818 The subclass shown above sets ``__slots__`` to an empty tuple. This keeps
819 keep memory requirements low by preventing the creation of instance dictionaries.
821 Subclassing is not useful for adding new, stored fields. Instead, simply
822 create a new named tuple type from the :attr:`_fields` attribute:
824 >>> Point3D = namedtuple('Point3D', Point._fields + ('z',))
826 Default values can be implemented by using :meth:`_replace` to
827 customize a prototype instance:
829 >>> Account = namedtuple('Account', 'owner balance transaction_count')
830 >>> default_account = Account('<owner name>', 0.0, 0)
831 >>> johns_account = default_account._replace(owner='John')
833 Enumerated constants can be implemented with named tuples, but it is simpler
834 and more efficient to use a simple class declaration:
836 >>> Status = namedtuple('Status', 'open pending closed')._make(range(3))
837 >>> Status.open, Status.pending, Status.closed
840 ... open, pending, closed = range(3)
844 `Named tuple recipe <http://code.activestate.com/recipes/500261/>`_
845 adapted for Python 2.4.
848 :class:`OrderedDict` objects
849 ----------------------------
851 Ordered dictionaries are just like regular dictionaries but they remember the
852 order that items were inserted. When iterating over an ordered dictionary,
853 the items are returned in the order their keys were first added.
855 .. class:: OrderedDict([items])
857 Return an instance of a dict subclass, supporting the usual :class:`dict`
858 methods. An *OrderedDict* is a dict that remembers the order that keys
859 were first inserted. If a new entry overwrites an existing entry, the
860 original insertion position is left unchanged. Deleting an entry and
861 reinserting it will move it to the end.
863 .. versionadded:: 2.7
865 .. method:: OrderedDict.popitem(last=True)
867 The :meth:`popitem` method for ordered dictionaries returns and removes
868 a (key, value) pair. The pairs are returned in LIFO order if *last* is
869 true or FIFO order if false.
871 Equality tests between :class:`OrderedDict` objects are order-sensitive
872 and are implemented as ``list(od1.items())==list(od2.items())``.
873 Equality tests between :class:`OrderedDict` objects and other
874 :class:`Mapping` objects are order-insensitive like regular dictionaries.
875 This allows :class:`OrderedDict` objects to be substituted anywhere a
876 regular dictionary is used.
878 The :class:`OrderedDict` constructor and :meth:`update` method both accept
879 keyword arguments, but their order is lost because Python's function call
880 semantics pass-in keyword arguments using a regular unordered dictionary.
884 `Equivalent OrderedDict recipe <http://code.activestate.com/recipes/576693/>`_
885 that runs on Python 2.4 or later.