^^&&^^

Edit File: datamodel.html

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN"
 "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">

<html xmlns="http://www.w3.org/1999/xhtml">
 <head>
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
 
 <title>3. Data model &mdash; Python 2.7.5 documentation</title>
 
 <link rel="stylesheet" href="../_static/default.css" type="text/css" />
 <link rel="stylesheet" href="../_static/pygments.css" type="text/css" />
 
 <script type="text/javascript">
 var DOCUMENTATION_OPTIONS = {
 URL_ROOT: '../',
 VERSION: '2.7.5',
 COLLAPSE_INDEX: false,
 FILE_SUFFIX: '.html',
 HAS_SOURCE: true
 };
 </script>
 <script type="text/javascript" src="../_static/jquery.js"></script>
 <script type="text/javascript" src="../_static/underscore.js"></script>
 <script type="text/javascript" src="../_static/doctools.js"></script>
 <script type="text/javascript" src="../_static/sidebar.js"></script>
 <link rel="search" type="application/opensearchdescription+xml"
 title="Search within Python 2.7.5 documentation"
 href="../_static/opensearch.xml"/>
 <link rel="author" title="About these documents" href="../about.html" />
 <link rel="copyright" title="Copyright" href="../copyright.html" />
 <link rel="top" title="Python 2.7.5 documentation" href="../index.html" />
 <link rel="up" title="The Python Language Reference" href="index.html" />
 <link rel="next" title="4. Execution model" href="executionmodel.html" />
 <link rel="prev" title="2. Lexical analysis" href="lexical_analysis.html" />
 <link rel="shortcut icon" type="image/png" href="../_static/py.png" />
 <script type="text/javascript" src="../_static/copybutton.js"></script>

</head>
 <body>
 <div class="related">
 <h3>Navigation</h3>
 <ul>
 <li class="right" style="margin-right: 10px">
 <a href="../genindex.html" title="General Index"
 accesskey="I">index</a></li>
 <li class="right" >
 <a href="../py-modindex.html" title="Python Module Index"
 >modules</a> |</li>
 <li class="right" >
 <a href="executionmodel.html" title="4. Execution model"
 accesskey="N">next</a> |</li>
 <li class="right" >
 <a href="lexical_analysis.html" title="2. Lexical analysis"
 accesskey="P">previous</a> |</li>
 <li><img src="../_static/py.png" alt=""
 style="vertical-align: middle; margin-top: -1px"/></li>
 <li><a href="http://www.python.org/">Python</a> &raquo;</li>
 <li>
 <a href="../index.html">Python 2.7.5 documentation</a> &raquo;
 </li>

<li><a href="index.html" accesskey="U">The Python Language Reference</a> &raquo;</li> 
 </ul>
 </div>

<div class="document">
 <div class="documentwrapper">
 <div class="bodywrapper">
 <div class="body">
 
 <div class="section" id="data-model">
<h1>3. Data model<a class="headerlink" href="#data-model" title="Permalink to this headline">¶</a></h1>
<div class="section" id="objects-values-and-types">
<h2>3.1. Objects, values and types<a class="headerlink" href="#objects-values-and-types" title="Permalink to this headline">¶</a></h2>
Objects are Python&#8217;s abstraction for data. All data in a Python program
is represented by objects or by relations between objects. (In a sense, and in
conformance to Von Neumann&#8217;s model of a &#8220;stored program computer,&#8221; code is also
represented by objects.)
Every object has an identity, a type and a value. An object&#8217;s identity never
changes once it has been created; you may think of it as the object&#8217;s address in
memory. The &#8216;<a class="reference internal" href="expressions.html#is"><tt class="xref std std-keyword docutils literal">is</tt></a>&#8216; operator compares the identity of two objects; the
<a class="reference internal" href="../library/functions.html#id" title="id"><tt class="xref py py-func docutils literal">id()</tt></a> function returns an integer representing its identity (currently
implemented as its address). An object&#8217;s type is also unchangeable. <a class="footnote-reference" href="#id5" id="id1">[1]</a>
An object&#8217;s type determines the operations that the object supports (e.g., &#8220;does
it have a length?&#8221;) and also defines the possible values for objects of that
type. The <a class="reference internal" href="../library/functions.html#type" title="type"><tt class="xref py py-func docutils literal">type()</tt></a> function returns an object&#8217;s type (which is an object
itself). The value of some objects can change. Objects whose value can
change are said to be mutable; objects whose value is unchangeable once they
are created are called immutable. (The value of an immutable container object
that contains a reference to a mutable object can change when the latter&#8217;s value
is changed; however the container is still considered immutable, because the
collection of objects it contains cannot be changed. So, immutability is not
strictly the same as having an unchangeable value, it is more subtle.) An
object&#8217;s mutability is determined by its type; for instance, numbers, strings
and tuples are immutable, while dictionaries and lists are mutable.
Objects are never explicitly destroyed; however, when they become unreachable
they may be garbage-collected. An implementation is allowed to postpone garbage
collection or omit it altogether &#8212; it is a matter of implementation quality
how garbage collection is implemented, as long as no objects are collected that
are still reachable.
<div class="impl-detail compound">
CPython implementation detail: CPython currently uses a reference-counting scheme with (optional) delayed
detection of cyclically linked garbage, which collects most objects as soon
as they become unreachable, but is not guaranteed to collect garbage
containing circular references. See the documentation of the <a class="reference internal" href="../library/gc.html#module-gc" title="gc: Interface to the cycle-detecting garbage collector."><tt class="xref py py-mod docutils literal">gc</tt></a>
module for information on controlling the collection of cyclic garbage.
Other implementations act differently and CPython may change.
Do not depend on immediate finalization of objects when they become
unreachable (ex: always close files).
</div>
Note that the use of the implementation&#8217;s tracing or debugging facilities may
keep objects alive that would normally be collectable. Also note that catching
an exception with a &#8216;<a class="reference internal" href="compound_stmts.html#try"><tt class="xref std std-keyword docutils literal">try</tt></a>...<a class="reference internal" href="compound_stmts.html#except"><tt class="xref std std-keyword docutils literal">except</tt></a>&#8216; statement may keep
objects alive.
Some objects contain references to &#8220;external&#8221; resources such as open files or
windows. It is understood that these resources are freed when the object is
garbage-collected, but since garbage collection is not guaranteed to happen,
such objects also provide an explicit way to release the external resource,
usually a <tt class="xref py py-meth docutils literal">close()</tt> method. Programs are strongly recommended to explicitly
close such objects. The &#8216;<a class="reference internal" href="compound_stmts.html#try"><tt class="xref std std-keyword docutils literal">try</tt></a>...<a class="reference internal" href="compound_stmts.html#finally"><tt class="xref std std-keyword docutils literal">finally</tt></a>&#8216; statement
provides a convenient way to do this.
Some objects contain references to other objects; these are called containers.
Examples of containers are tuples, lists and dictionaries. The references are
part of a container&#8217;s value. In most cases, when we talk about the value of a
container, we imply the values, not the identities of the contained objects;
however, when we talk about the mutability of a container, only the identities
of the immediately contained objects are implied. So, if an immutable container
(like a tuple) contains a reference to a mutable object, its value changes if
that mutable object is changed.
Types affect almost all aspects of object behavior. Even the importance of
object identity is affected in some sense: for immutable types, operations that
compute new values may actually return a reference to any existing object with
the same type and value, while for mutable objects this is not allowed. E.g.,
after <tt class="docutils literal">a = 1; b = 1</tt>, <tt class="docutils literal">a</tt> and <tt class="docutils literal">b</tt> may or may not refer to the same object
with the value one, depending on the implementation, but after <tt class="docutils literal">c = []; d =
[]</tt>, <tt class="docutils literal">c</tt> and <tt class="docutils literal">d</tt> are guaranteed to refer to two different, unique, newly
created empty lists. (Note that <tt class="docutils literal">c = d = []</tt> assigns the same object to both
<tt class="docutils literal">c</tt> and <tt class="docutils literal">d</tt>.)
</div>
<div class="section" id="the-standard-type-hierarchy">
<h2>3.2. The standard type hierarchy<a class="headerlink" href="#the-standard-type-hierarchy" title="Permalink to this headline">¶</a></h2>
Below is a list of the types that are built into Python. Extension modules
(written in C, Java, or other languages, depending on the implementation) can
define additional types. Future versions of Python may add types to the type
hierarchy (e.g., rational numbers, efficiently stored arrays of integers, etc.).
Some of the type descriptions below contain a paragraph listing &#8216;special
attributes.&#8217; These are attributes that provide access to the implementation and
are not intended for general use. Their definition may change in the future.
<dl class="docutils">
<dt>None</dt>
<dd>This type has a single value. There is a single object with this value. This
object is accessed through the built-in name <tt class="docutils literal">None</tt>. It is used to signify the
absence of a value in many situations, e.g., it is returned from functions that
don&#8217;t explicitly return anything. Its truth value is false.
</dd>
<dt>NotImplemented</dt>
<dd>This type has a single value. There is a single object with this value. This
object is accessed through the built-in name <tt class="docutils literal">NotImplemented</tt>. Numeric methods
and rich comparison methods may return this value if they do not implement the
operation for the operands provided. (The interpreter will then try the
reflected operation, or some other fallback, depending on the operator.) Its
truth value is true.
</dd>
<dt>Ellipsis</dt>
<dd>This type has a single value. There is a single object with this value. This
object is accessed through the built-in name <tt class="docutils literal">Ellipsis</tt>. It is used to
indicate the presence of the <tt class="docutils literal">...</tt> syntax in a slice. Its truth value is
true.
</dd>
<dt><a class="reference internal" href="../library/numbers.html#numbers.Number" title="numbers.Number"><tt class="xref py py-class docutils literal">numbers.Number</tt></a></dt>
<dd>These are created by numeric literals and returned as results by arithmetic
operators and arithmetic built-in functions. Numeric objects are immutable;
once created their value never changes. Python numbers are of course strongly
related to mathematical numbers, but subject to the limitations of numerical
representation in computers.
Python distinguishes between integers, floating point numbers, and complex
numbers:
<dl class="last docutils">
<dt><a class="reference internal" href="../library/numbers.html#numbers.Integral" title="numbers.Integral"><tt class="xref py py-class docutils literal">numbers.Integral</tt></a></dt>
<dd>These represent elements from the mathematical set of integers (positive and
negative).
There are three types of integers:
<dl class="docutils">
<dt>Plain integers</dt>
<dd>These represent numbers in the range -2147483648 through 2147483647.
(The range may be larger on machines with a larger natural word size,
but not smaller.) When the result of an operation would fall outside
this range, the result is normally returned as a long integer (in some
cases, the exception <a class="reference internal" href="../library/exceptions.html#exceptions.OverflowError" title="exceptions.OverflowError"><tt class="xref py py-exc docutils literal">OverflowError</tt></a> is raised instead). For the
purpose of shift and mask operations, integers are assumed to have a
binary, 2&#8217;s complement notation using 32 or more bits, and hiding no
bits from the user (i.e., all 4294967296 different bit patterns
correspond to different values).
</dd>
<dt>Long integers</dt>
<dd>These represent numbers in an unlimited range, subject to available
(virtual) memory only. For the purpose of shift and mask operations, a
binary representation is assumed, and negative numbers are represented
in a variant of 2&#8217;s complement which gives the illusion of an infinite
string of sign bits extending to the left.
</dd>
<dt>Booleans</dt>
<dd>These represent the truth values False and True. The two objects
representing the values False and True are the only Boolean objects.
The Boolean type is a subtype of plain integers, and Boolean values
behave like the values 0 and 1, respectively, in almost all contexts,
the exception being that when converted to a string, the strings
<tt class="docutils literal">&quot;False&quot;</tt> or <tt class="docutils literal">&quot;True&quot;</tt> are returned, respectively.
</dd>
</dl>
The rules for integer representation are intended to give the most
meaningful interpretation of shift and mask operations involving negative
integers and the least surprises when switching between the plain and long
integer domains. Any operation, if it yields a result in the plain
integer domain, will yield the same result in the long integer domain or
when using mixed operands. The switch between domains is transparent to
the programmer.
</dd>
<dt><a class="reference internal" href="../library/numbers.html#numbers.Real" title="numbers.Real"><tt class="xref py py-class docutils literal">numbers.Real</tt></a> (<a class="reference internal" href="../library/functions.html#float" title="float"><tt class="xref py py-class docutils literal">float</tt></a>)</dt>
<dd>These represent machine-level double precision floating point numbers. You are
at the mercy of the underlying machine architecture (and C or Java
implementation) for the accepted range and handling of overflow. Python does not
support single-precision floating point numbers; the savings in processor and
memory usage that are usually the reason for using these is dwarfed by the
overhead of using objects in Python, so there is no reason to complicate the
language with two kinds of floating point numbers.
</dd>
<dt><a class="reference internal" href="../library/numbers.html#numbers.Complex" title="numbers.Complex"><tt class="xref py py-class docutils literal">numbers.Complex</tt></a></dt>
<dd>These represent complex numbers as a pair of machine-level double precision
floating point numbers. The same caveats apply as for floating point numbers.
The real and imaginary parts of a complex number <tt class="docutils literal">z</tt> can be retrieved through
the read-only attributes <tt class="docutils literal">z.real</tt> and <tt class="docutils literal">z.imag</tt>.
</dd>
</dl>
</dd>
<dt>Sequences</dt>
<dd>These represent finite ordered sets indexed by non-negative numbers. The
built-in function <a class="reference internal" href="../library/functions.html#len" title="len"><tt class="xref py py-func docutils literal">len()</tt></a> returns the number of items of a sequence. When
the length of a sequence is n, the index set contains the numbers 0, 1,
..., n-1. Item i of sequence a is selected by <tt class="docutils literal">a[i]</tt>.
Sequences also support slicing: <tt class="docutils literal">a[i:j]</tt> selects all items with index k such
that i <tt class="docutils literal">&lt;=</tt> k <tt class="docutils literal">&lt;</tt> j. When used as an expression, a slice is a
sequence of the same type. This implies that the index set is renumbered so
that it starts at 0.
Some sequences also support &#8220;extended slicing&#8221; with a third &#8220;step&#8221; parameter:
<tt class="docutils literal">a[i:j:k]</tt> selects all items of a with index x where <tt class="docutils literal">x = i + n*k</tt>, n
<tt class="docutils literal">&gt;=</tt> <tt class="docutils literal">0</tt> and i <tt class="docutils literal">&lt;=</tt> x <tt class="docutils literal">&lt;</tt> j.
Sequences are distinguished according to their mutability:
<dl class="last docutils">
<dt>Immutable sequences</dt>
<dd>An object of an immutable sequence type cannot change once it is created. (If
the object contains references to other objects, these other objects may be
mutable and may be changed; however, the collection of objects directly
referenced by an immutable object cannot change.)
The following types are immutable sequences:
<dl class="last docutils">
<dt>Strings</dt>
<dd>The items of a string are characters. There is no separate character type; a
character is represented by a string of one item. Characters represent (at
least) 8-bit bytes. The built-in functions <a class="reference internal" href="../library/functions.html#chr" title="chr"><tt class="xref py py-func docutils literal">chr()</tt></a> and <a class="reference internal" href="../library/functions.html#ord" title="ord"><tt class="xref py py-func docutils literal">ord()</tt></a> convert
between characters and nonnegative integers representing the byte values. Bytes
with the values 0-127 usually represent the corresponding ASCII values, but the
interpretation of values is up to the program. The string data type is also
used to represent arrays of bytes, e.g., to hold data read from a file.
(On systems whose native character set is not ASCII, strings may use EBCDIC in
their internal representation, provided the functions <a class="reference internal" href="../library/functions.html#chr" title="chr"><tt class="xref py py-func docutils literal">chr()</tt></a> and
<a class="reference internal" href="../library/functions.html#ord" title="ord"><tt class="xref py py-func docutils literal">ord()</tt></a> implement a mapping between ASCII and EBCDIC, and string comparison
preserves the ASCII order. Or perhaps someone can propose a better rule?)
</dd>
<dt>Unicode</dt>
<dd>The items of a Unicode object are Unicode code units. A Unicode code unit is
represented by a Unicode object of one item and can hold either a 16-bit or
32-bit value representing a Unicode ordinal (the maximum value for the ordinal
is given in <tt class="docutils literal">sys.maxunicode</tt>, and depends on how Python is configured at
compile time). Surrogate pairs may be present in the Unicode object, and will
be reported as two separate items. The built-in functions <a class="reference internal" href="../library/functions.html#unichr" title="unichr"><tt class="xref py py-func docutils literal">unichr()</tt></a> and
<a class="reference internal" href="../library/functions.html#ord" title="ord"><tt class="xref py py-func docutils literal">ord()</tt></a> convert between code units and nonnegative integers representing the
Unicode ordinals as defined in the Unicode Standard 3.0. Conversion from and to
other encodings are possible through the Unicode method <tt class="xref py py-meth docutils literal">encode()</tt> and the
built-in function <a class="reference internal" href="../library/functions.html#unicode" title="unicode"><tt class="xref py py-func docutils literal">unicode()</tt></a>.
</dd>
<dt>Tuples</dt>
<dd>The items of a tuple are arbitrary Python objects. Tuples of two or more items
are formed by comma-separated lists of expressions. A tuple of one item (a
&#8216;singleton&#8217;) can be formed by affixing a comma to an expression (an expression
by itself does not create a tuple, since parentheses must be usable for grouping
of expressions). An empty tuple can be formed by an empty pair of parentheses.
</dd>
</dl>
</dd>
<dt>Mutable sequences</dt>
<dd>Mutable sequences can be changed after they are created. The subscription and
slicing notations can be used as the target of assignment and <a class="reference internal" href="simple_stmts.html#del"><tt class="xref std std-keyword docutils literal">del</tt></a>
(delete) statements.
There are currently two intrinsic mutable sequence types:
<dl class="docutils">
<dt>Lists</dt>
<dd>The items of a list are arbitrary Python objects. Lists are formed by placing a
comma-separated list of expressions in square brackets. (Note that there are no
special cases needed to form lists of length 0 or 1.)
</dd>
<dt>Byte Arrays</dt>
<dd>A bytearray object is a mutable array. They are created by the built-in
<a class="reference internal" href="../library/functions.html#bytearray" title="bytearray"><tt class="xref py py-func docutils literal">bytearray()</tt></a> constructor. Aside from being mutable (and hence
unhashable), byte arrays otherwise provide the same interface and
functionality as immutable bytes objects.
</dd>
</dl>
The extension module <a class="reference internal" href="../library/array.html#module-array" title="array: Space efficient arrays of uniformly typed numeric values."><tt class="xref py py-mod docutils literal">array</tt></a> provides an additional example of a mutable
sequence type.
</dd>
</dl>
</dd>
<dt>Set types</dt>
<dd>These represent unordered, finite sets of unique, immutable objects. As such,
they cannot be indexed by any subscript. However, they can be iterated over, and
the built-in function <a class="reference internal" href="../library/functions.html#len" title="len"><tt class="xref py py-func docutils literal">len()</tt></a> returns the number of items in a set. Common
uses for sets are fast membership testing, removing duplicates from a sequence,
and computing mathematical operations such as intersection, union, difference,
and symmetric difference.
For set elements, the same immutability rules apply as for dictionary keys. Note
that numeric types obey the normal rules for numeric comparison: if two numbers
compare equal (e.g., <tt class="docutils literal">1</tt> and <tt class="docutils literal">1.0</tt>), only one of them can be contained in a
set.
There are currently two intrinsic set types:
<dl class="last docutils">
<dt>Sets</dt>
<dd>These represent a mutable set. They are created by the built-in <a class="reference internal" href="../library/stdtypes.html#set" title="set"><tt class="xref py py-func docutils literal">set()</tt></a>
constructor and can be modified afterwards by several methods, such as
<tt class="xref py py-meth docutils literal">add()</tt>.
</dd>
<dt>Frozen sets</dt>
<dd>These represent an immutable set. They are created by the built-in
<a class="reference internal" href="../library/stdtypes.html#frozenset" title="frozenset"><tt class="xref py py-func docutils literal">frozenset()</tt></a> constructor. As a frozenset is immutable and
<a class="reference internal" href="../glossary.html#term-hashable">hashable</a>, it can be used again as an element of another set, or as
a dictionary key.
</dd>
</dl>
</dd>
<dt>Mappings</dt>
<dd>These represent finite sets of objects indexed by arbitrary index sets. The
subscript notation <tt class="docutils literal">a[k]</tt> selects the item indexed by <tt class="docutils literal">k</tt> from the mapping
<tt class="docutils literal">a</tt>; this can be used in expressions and as the target of assignments or
<a class="reference internal" href="simple_stmts.html#del"><tt class="xref std std-keyword docutils literal">del</tt></a> statements. The built-in function <a class="reference internal" href="../library/functions.html#len" title="len"><tt class="xref py py-func docutils literal">len()</tt></a> returns the number
of items in a mapping.
There is currently a single intrinsic mapping type:
<dl class="last docutils">
<dt>Dictionaries</dt>
<dd>These represent finite sets of objects indexed by nearly arbitrary values. The
only types of values not acceptable as keys are values containing lists or
dictionaries or other mutable types that are compared by value rather than by
object identity, the reason being that the efficient implementation of
dictionaries requires a key&#8217;s hash value to remain constant. Numeric types used
for keys obey the normal rules for numeric comparison: if two numbers compare
equal (e.g., <tt class="docutils literal">1</tt> and <tt class="docutils literal">1.0</tt>) then they can be used interchangeably to index
the same dictionary entry.
Dictionaries are mutable; they can be created by the <tt class="docutils literal">{...}</tt> notation (see
section <a class="reference internal" href="expressions.html#dict">Dictionary displays</a>).
The extension modules <a class="reference internal" href="../library/dbm.html#module-dbm" title="dbm: The standard &quot;database&quot; interface, based on ndbm. (Unix)"><tt class="xref py py-mod docutils literal">dbm</tt></a>, <a class="reference internal" href="../library/gdbm.html#module-gdbm" title="gdbm: GNU's reinterpretation of dbm. (Unix)"><tt class="xref py py-mod docutils literal">gdbm</tt></a>, and <a class="reference internal" href="../library/bsddb.html#module-bsddb" title="bsddb: Interface to Berkeley DB database library"><tt class="xref py py-mod docutils literal">bsddb</tt></a> provide
additional examples of mapping types.
</dd>
</dl>
</dd>
<dt>Callable types</dt>
<dd>These are the types to which the function call operation (see section
<a class="reference internal" href="expressions.html#calls">Calls</a>) can be applied:
<dl class="last docutils">
<dt>User-defined functions</dt>
<dd>A user-defined function object is created by a function definition (see
section <a class="reference internal" href="compound_stmts.html#function">Function definitions</a>). It should be called with an argument list
containing the same number of items as the function&#8217;s formal parameter
list.
Special attributes:
<table border="1" class="docutils">
<colgroup>
<col width="35%" />
<col width="48%" />
<col width="17%" />
</colgroup>
<thead valign="bottom">
<tr class="row-odd"><th class="head">Attribute</th>
<th class="head">Meaning</th>
<th class="head">&nbsp;</th>
</tr>
</thead>
<tbody valign="top">
<tr class="row-even"><td><tt class="xref py py-attr docutils literal">func_doc</tt></td>
<td>The function&#8217;s documentation
string, or <tt class="docutils literal">None</tt> if
unavailable</td>
<td>Writable</td>
</tr>
<tr class="row-odd"><td><tt class="xref py py-attr docutils literal">__doc__</tt></td>
<td>Another way of spelling
<tt class="xref py py-attr docutils literal">func_doc</tt></td>
<td>Writable</td>
</tr>
<tr class="row-even"><td><tt class="xref py py-attr docutils literal">func_name</tt></td>
<td>The function&#8217;s name</td>
<td>Writable</td>
</tr>
<tr class="row-odd"><td><tt class="xref py py-attr docutils literal">__name__</tt></td>
<td>Another way of spelling
<tt class="xref py py-attr docutils literal">func_name</tt></td>
<td>Writable</td>
</tr>
<tr class="row-even"><td><tt class="xref py py-attr docutils literal">__module__</tt></td>
<td>The name of the module the
function was defined in, or
<tt class="docutils literal">None</tt> if unavailable.</td>
<td>Writable</td>
</tr>
<tr class="row-odd"><td><tt class="xref py py-attr docutils literal">func_defaults</tt></td>
<td>A tuple containing default
argument values for those
arguments that have defaults,
or <tt class="docutils literal">None</tt> if no arguments
have a default value</td>
<td>Writable</td>
</tr>
<tr class="row-even"><td><tt class="xref py py-attr docutils literal">func_code</tt></td>
<td>The code object representing
the compiled function body.</td>
<td>Writable</td>
</tr>
<tr class="row-odd"><td><tt class="xref py py-attr docutils literal">func_globals</tt></td>
<td>A reference to the dictionary
that holds the function&#8217;s
global variables &#8212; the
global namespace of the
module in which the function
was defined.</td>
<td>Read-only</td>
</tr>
<tr class="row-even"><td><tt class="xref py py-attr docutils literal">func_dict</tt></td>
<td>The namespace supporting
arbitrary function
attributes.</td>
<td>Writable</td>
</tr>
<tr class="row-odd"><td><tt class="xref py py-attr docutils literal">func_closure</tt></td>
<td><tt class="docutils literal">None</tt> or a tuple of cells
that contain bindings for the
function&#8217;s free variables.</td>
<td>Read-only</td>
</tr>
</tbody>
</table>
Most of the attributes labelled &#8220;Writable&#8221; check the type of the assigned value.

Changed in version 2.4: <tt class="docutils literal">func_name</tt> is now writable.
Function objects also support getting and setting arbitrary attributes, which
can be used, for example, to attach metadata to functions. Regular attribute
dot-notation is used to get and set such attributes. Note that the current
implementation only supports function attributes on user-defined functions.
Function attributes on built-in functions may be supported in the future.
Additional information about a function&#8217;s definition can be retrieved from its
code object; see the description of internal types below.
</dd>
<dt>User-defined methods</dt>
<dd>A user-defined method object combines a class, a class instance (or <tt class="docutils literal">None</tt>)
and any callable object (normally a user-defined function).
Special read-only attributes: <tt class="xref py py-attr docutils literal">im_self</tt> is the class instance object,
<tt class="xref py py-attr docutils literal">im_func</tt> is the function object; <tt class="xref py py-attr docutils literal">im_class</tt> is the class of
<tt class="xref py py-attr docutils literal">im_self</tt> for bound methods or the class that asked for the method for
unbound methods; <tt class="xref py py-attr docutils literal">__doc__</tt> is the method&#8217;s documentation (same as
<tt class="docutils literal">im_func.__doc__</tt>); <tt class="xref py py-attr docutils literal">__name__</tt> is the method name (same as
<tt class="docutils literal">im_func.__name__</tt>); <tt class="xref py py-attr docutils literal">__module__</tt> is the name of the module the method
was defined in, or <tt class="docutils literal">None</tt> if unavailable.

Changed in version 2.2: <tt class="xref py py-attr docutils literal">im_self</tt> used to refer to the class that defined the method.

Changed in version 2.6: For Python 3 forward-compatibility, <tt class="xref py py-attr docutils literal">im_func</tt> is also available as
<tt class="xref py py-attr docutils literal">__func__</tt>, and <tt class="xref py py-attr docutils literal">im_self</tt> as <tt class="xref py py-attr docutils literal">__self__</tt>.
Methods also support accessing (but not setting) the arbitrary function
attributes on the underlying function object.
User-defined method objects may be created when getting an attribute of a class
(perhaps via an instance of that class), if that attribute is a user-defined
function object, an unbound user-defined method object, or a class method
object. When the attribute is a user-defined method object, a new method object
is only created if the class from which it is being retrieved is the same as, or
a derived class of, the class stored in the original method object; otherwise,
the original method object is used as it is.
When a user-defined method object is created by retrieving a user-defined
function object from a class, its <tt class="xref py py-attr docutils literal">im_self</tt> attribute is <tt class="docutils literal">None</tt>
and the method object is said to be unbound. When one is created by
retrieving a user-defined function object from a class via one of its
instances, its <tt class="xref py py-attr docutils literal">im_self</tt> attribute is the instance, and the method
object is said to be bound. In either case, the new method&#8217;s
<tt class="xref py py-attr docutils literal">im_class</tt> attribute is the class from which the retrieval takes
place, and its <tt class="xref py py-attr docutils literal">im_func</tt> attribute is the original function object.
When a user-defined method object is created by retrieving another method object
from a class or instance, the behaviour is the same as for a function object,
except that the <tt class="xref py py-attr docutils literal">im_func</tt> attribute of the new instance is not the
original method object but its <tt class="xref py py-attr docutils literal">im_func</tt> attribute.
When a user-defined method object is created by retrieving a class method object
from a class or instance, its <tt class="xref py py-attr docutils literal">im_self</tt> attribute is the class itself, and
its <tt class="xref py py-attr docutils literal">im_func</tt> attribute is the function object underlying the class method.
When an unbound user-defined method object is called, the underlying function
(<tt class="xref py py-attr docutils literal">im_func</tt>) is called, with the restriction that the first argument must
be an instance of the proper class (<tt class="xref py py-attr docutils literal">im_class</tt>) or of a derived class
thereof.
When a bound user-defined method object is called, the underlying function
(<tt class="xref py py-attr docutils literal">im_func</tt>) is called, inserting the class instance (<tt class="xref py py-attr docutils literal">im_self</tt>) in
front of the argument list. For instance, when <tt class="xref py py-class docutils literal">C</tt> is a class which
contains a definition for a function <tt class="xref py py-meth docutils literal">f()</tt>, and <tt class="docutils literal">x</tt> is an instance of
<tt class="xref py py-class docutils literal">C</tt>, calling <tt class="docutils literal">x.f(1)</tt> is equivalent to calling <tt class="docutils literal">C.f(x, 1)</tt>.
When a user-defined method object is derived from a class method object, the
&#8220;class instance&#8221; stored in <tt class="xref py py-attr docutils literal">im_self</tt> will actually be the class itself, so
that calling either <tt class="docutils literal">x.f(1)</tt> or <tt class="docutils literal">C.f(1)</tt> is equivalent to calling <tt class="docutils literal">f(C,1)</tt>
where <tt class="docutils literal">f</tt> is the underlying function.
Note that the transformation from function object to (unbound or bound) method
object happens each time the attribute is retrieved from the class or instance.
In some cases, a fruitful optimization is to assign the attribute to a local
variable and call that local variable. Also notice that this transformation only
happens for user-defined functions; other callable objects (and all non-callable
objects) are retrieved without transformation. It is also important to note
that user-defined functions which are attributes of a class instance are not
converted to bound methods; this only happens when the function is an
attribute of the class.
</dd>
<dt>Generator functions</dt>
<dd>A function or method which uses the <a class="reference internal" href="simple_stmts.html#yield"><tt class="xref std std-keyword docutils literal">yield</tt></a> statement (see section
<a class="reference internal" href="simple_stmts.html#yield">The yield statement</a>) is called a generator
function. Such a function, when called, always returns an iterator object
which can be used to execute the body of the function: calling the iterator&#8217;s
<a class="reference internal" href="../library/functions.html#next" title="next"><tt class="xref py py-meth docutils literal">next()</tt></a> method will cause the function to execute until it provides a value
using the <a class="reference internal" href="simple_stmts.html#yield"><tt class="xref std std-keyword docutils literal">yield</tt></a> statement. When the function executes a
<a class="reference internal" href="simple_stmts.html#return"><tt class="xref std std-keyword docutils literal">return</tt></a> statement or falls off the end, a <a class="reference internal" href="../library/exceptions.html#exceptions.StopIteration" title="exceptions.StopIteration"><tt class="xref py py-exc docutils literal">StopIteration</tt></a>
exception is raised and the iterator will have reached the end of the set of
values to be returned.
</dd>
<dt>Built-in functions</dt>
<dd>A built-in function object is a wrapper around a C function. Examples of
built-in functions are <a class="reference internal" href="../library/functions.html#len" title="len"><tt class="xref py py-func docutils literal">len()</tt></a> and <a class="reference internal" href="../library/math.html#math.sin" title="math.sin"><tt class="xref py py-func docutils literal">math.sin()</tt></a> (<a class="reference internal" href="../library/math.html#module-math" title="math: Mathematical functions (sin() etc.)."><tt class="xref py py-mod docutils literal">math</tt></a> is a
standard built-in module). The number and type of the arguments are
determined by the C function. Special read-only attributes:
<tt class="xref py py-attr docutils literal">__doc__</tt> is the function&#8217;s documentation string, or <tt class="docutils literal">None</tt> if
unavailable; <tt class="xref py py-attr docutils literal">__name__</tt> is the function&#8217;s name; <tt class="xref py py-attr docutils literal">__self__</tt> is
set to <tt class="docutils literal">None</tt> (but see the next item); <tt class="xref py py-attr docutils literal">__module__</tt> is the name of
the module the function was defined in or <tt class="docutils literal">None</tt> if unavailable.
</dd>
<dt>Built-in methods</dt>
<dd>This is really a different disguise of a built-in function, this time containing
an object passed to the C function as an implicit extra argument. An example of
a built-in method is <tt class="docutils literal">alist.append()</tt>, assuming alist is a list object. In
this case, the special read-only attribute <tt class="xref py py-attr docutils literal">__self__</tt> is set to the object
denoted by alist.
</dd>
<dt>Class Types</dt>
<dd>Class types, or &#8220;new-style classes,&#8221; are callable. These objects normally act
as factories for new instances of themselves, but variations are possible for
class types that override <a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a>. The arguments of the call are passed
to <a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a> and, in the typical case, to <a class="reference internal" href="#object.__init__" title="object.__init__"><tt class="xref py py-meth docutils literal">__init__()</tt></a> to initialize
the new instance.</dd>
<dt>Classic Classes</dt>
<dd>Class objects are described below. When a class object is called, a new class
instance (also described below) is created and returned. This implies a call to
the class&#8217;s <a class="reference internal" href="#object.__init__" title="object.__init__"><tt class="xref py py-meth docutils literal">__init__()</tt></a> method if it has one. Any arguments are passed on
to the <a class="reference internal" href="#object.__init__" title="object.__init__"><tt class="xref py py-meth docutils literal">__init__()</tt></a> method. If there is no <a class="reference internal" href="#object.__init__" title="object.__init__"><tt class="xref py py-meth docutils literal">__init__()</tt></a> method, the
class must be called without arguments.
</dd>
<dt>Class instances</dt>
<dd>Class instances are described below. Class instances are callable only when the
class has a <a class="reference internal" href="#object.__call__" title="object.__call__"><tt class="xref py py-meth docutils literal">__call__()</tt></a> method; <tt class="docutils literal">x(arguments)</tt> is a shorthand for
<tt class="docutils literal">x.__call__(arguments)</tt>.</dd>
</dl>
</dd>
<dt>Modules</dt>
<dd>Modules are imported by the <a class="reference internal" href="simple_stmts.html#import"><tt class="xref std std-keyword docutils literal">import</tt></a> statement (see section
<a class="reference internal" href="simple_stmts.html#import">The import statement</a>). A module object has a
namespace implemented by a dictionary object (this is the dictionary referenced
by the func_globals attribute of functions defined in the module). Attribute
references are translated to lookups in this dictionary, e.g., <tt class="docutils literal">m.x</tt> is
equivalent to <tt class="docutils literal">m.__dict__[&quot;x&quot;]</tt>. A module object does not contain the code
object used to initialize the module (since it isn&#8217;t needed once the
initialization is done).
Attribute assignment updates the module&#8217;s namespace dictionary, e.g., <tt class="docutils literal">m.x =
1</tt> is equivalent to <tt class="docutils literal">m.__dict__[&quot;x&quot;] = 1</tt>.
Special read-only attribute: <tt class="xref py py-attr docutils literal">__dict__</tt> is the module&#8217;s namespace as a
dictionary object.
<div class="impl-detail compound">
CPython implementation detail: Because of the way CPython clears module dictionaries, the module
dictionary will be cleared when the module falls out of scope even if the
dictionary still has live references. To avoid this, copy the dictionary
or keep the module around while using its dictionary directly.
</div>
Predefined (writable) attributes: <tt class="xref py py-attr docutils literal">__name__</tt> is the module&#8217;s name;
<tt class="xref py py-attr docutils literal">__doc__</tt> is the module&#8217;s documentation string, or <tt class="docutils literal">None</tt> if
unavailable; <tt class="xref py py-attr docutils literal">__file__</tt> is the pathname of the file from which the module
was loaded, if it was loaded from a file. The <tt class="xref py py-attr docutils literal">__file__</tt> attribute is not
present for C modules that are statically linked into the interpreter; for
extension modules loaded dynamically from a shared library, it is the pathname
of the shared library file.
</dd>
<dt>Classes</dt>
<dd>Both class types (new-style classes) and class objects (old-style/classic
classes) are typically created by class definitions (see section
<a class="reference internal" href="compound_stmts.html#class">Class definitions</a>). A class has a namespace implemented by a dictionary object.
Class attribute references are translated to lookups in this dictionary, e.g.,
<tt class="docutils literal">C.x</tt> is translated to <tt class="docutils literal">C.__dict__[&quot;x&quot;]</tt> (although for new-style classes
in particular there are a number of hooks which allow for other means of
locating attributes). When the attribute name is not found there, the
attribute search continues in the base classes. For old-style classes, the
search is depth-first, left-to-right in the order of occurrence in the base
class list. New-style classes use the more complex C3 method resolution
order which behaves correctly even in the presence of &#8216;diamond&#8217;
inheritance structures where there are multiple inheritance paths
leading back to a common ancestor. Additional details on the C3 MRO used by
new-style classes can be found in the documentation accompanying the
2.3 release at <a class="reference external" href="http://www.python.org/download/releases/2.3/mro/">http://www.python.org/download/releases/2.3/mro/</a>.
When a class attribute reference (for class <tt class="xref py py-class docutils literal">C</tt>, say) would yield a
user-defined function object or an unbound user-defined method object whose
associated class is either <tt class="xref py py-class docutils literal">C</tt> or one of its base classes, it is
transformed into an unbound user-defined method object whose <tt class="xref py py-attr docutils literal">im_class</tt>
attribute is <tt class="xref py py-class docutils literal">C</tt>. When it would yield a class method object, it is
transformed into a bound user-defined method object whose
<tt class="xref py py-attr docutils literal">im_self</tt> attribute is <tt class="xref py py-class docutils literal">C</tt>. When it would yield a
static method object, it is transformed into the object wrapped by the static
method object. See section <a class="reference internal" href="#descriptors">Implementing Descriptors</a> for another way in which
attributes retrieved from a class may differ from those actually contained in
its <tt class="xref py py-attr docutils literal">__dict__</tt> (note that only new-style classes support descriptors).
Class attribute assignments update the class&#8217;s dictionary, never the dictionary
of a base class.
A class object can be called (see above) to yield a class instance (see below).
Special attributes: <tt class="xref py py-attr docutils literal">__name__</tt> is the class name; <tt class="xref py py-attr docutils literal">__module__</tt> is
the module name in which the class was defined; <tt class="xref py py-attr docutils literal">__dict__</tt> is the
dictionary containing the class&#8217;s namespace; <tt class="xref py py-attr docutils literal">__bases__</tt> is a tuple
(possibly empty or a singleton) containing the base classes, in the order of
their occurrence in the base class list; <tt class="xref py py-attr docutils literal">__doc__</tt> is the class&#8217;s
documentation string, or None if undefined.
</dd>
<dt>Class instances</dt>
<dd>A class instance is created by calling a class object (see above). A class
instance has a namespace implemented as a dictionary which is the first place in
which attribute references are searched. When an attribute is not found there,
and the instance&#8217;s class has an attribute by that name, the search continues
with the class attributes. If a class attribute is found that is a user-defined
function object or an unbound user-defined method object whose associated class
is the class (call it <tt class="xref py py-class docutils literal">C</tt>) of the instance for which the attribute
reference was initiated or one of its bases, it is transformed into a bound
user-defined method object whose <tt class="xref py py-attr docutils literal">im_class</tt> attribute is <tt class="xref py py-class docutils literal">C</tt> and
whose <tt class="xref py py-attr docutils literal">im_self</tt> attribute is the instance. Static method and class method
objects are also transformed, as if they had been retrieved from class
<tt class="xref py py-class docutils literal">C</tt>; see above under &#8220;Classes&#8221;. See section <a class="reference internal" href="#descriptors">Implementing Descriptors</a> for
another way in which attributes of a class retrieved via its instances may
differ from the objects actually stored in the class&#8217;s <tt class="xref py py-attr docutils literal">__dict__</tt>. If no
class attribute is found, and the object&#8217;s class has a <a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><tt class="xref py py-meth docutils literal">__getattr__()</tt></a>
method, that is called to satisfy the lookup.
Attribute assignments and deletions update the instance&#8217;s dictionary, never a
class&#8217;s dictionary. If the class has a <a class="reference internal" href="#object.__setattr__" title="object.__setattr__"><tt class="xref py py-meth docutils literal">__setattr__()</tt></a> or
<a class="reference internal" href="#object.__delattr__" title="object.__delattr__"><tt class="xref py py-meth docutils literal">__delattr__()</tt></a> method, this is called instead of updating the instance
dictionary directly.
Class instances can pretend to be numbers, sequences, or mappings if they have
methods with certain special names. See section <a class="reference internal" href="#specialnames">Special method names</a>.
Special attributes: <tt class="xref py py-attr docutils literal">__dict__</tt> is the attribute dictionary;
<tt class="xref py py-attr docutils literal">__class__</tt> is the instance&#8217;s class.
</dd>
<dt>Files</dt>
<dd>A file object represents an open file. File objects are created by the
<a class="reference internal" href="../library/functions.html#open" title="open"><tt class="xref py py-func docutils literal">open()</tt></a> built-in function, and also by <a class="reference internal" href="../library/os.html#os.popen" title="os.popen"><tt class="xref py py-func docutils literal">os.popen()</tt></a>,
<a class="reference internal" href="../library/os.html#os.fdopen" title="os.fdopen"><tt class="xref py py-func docutils literal">os.fdopen()</tt></a>, and the <tt class="xref py py-meth docutils literal">makefile()</tt> method of socket objects (and
perhaps by other functions or methods provided by extension modules). The
objects <tt class="docutils literal">sys.stdin</tt>, <tt class="docutils literal">sys.stdout</tt> and <tt class="docutils literal">sys.stderr</tt> are initialized to
file objects corresponding to the interpreter&#8217;s standard input, output and
error streams. See <a class="reference internal" href="../library/stdtypes.html#bltin-file-objects">File Objects</a> for complete documentation of
file objects.
</dd>
<dt>Internal types</dt>
<dd>A few types used internally by the interpreter are exposed to the user. Their
definitions may change with future versions of the interpreter, but they are
mentioned here for completeness.
<dl class="docutils">
<dt>Code objects</dt>
<dd>Code objects represent byte-compiled executable Python code, or <a class="reference internal" href="../glossary.html#term-bytecode">bytecode</a>.
The difference between a code object and a function object is that the function
object contains an explicit reference to the function&#8217;s globals (the module in
which it was defined), while a code object contains no context; also the default
argument values are stored in the function object, not in the code object
(because they represent values calculated at run-time). Unlike function
objects, code objects are immutable and contain no references (directly or
indirectly) to mutable objects.
Special read-only attributes: <tt class="xref py py-attr docutils literal">co_name</tt> gives the function name;
<tt class="xref py py-attr docutils literal">co_argcount</tt> is the number of positional arguments (including arguments
with default values); <tt class="xref py py-attr docutils literal">co_nlocals</tt> is the number of local variables used
by the function (including arguments); <tt class="xref py py-attr docutils literal">co_varnames</tt> is a tuple containing
the names of the local variables (starting with the argument names);
<tt class="xref py py-attr docutils literal">co_cellvars</tt> is a tuple containing the names of local variables that are
referenced by nested functions; <tt class="xref py py-attr docutils literal">co_freevars</tt> is a tuple containing the
names of free variables; <tt class="xref py py-attr docutils literal">co_code</tt> is a string representing the sequence
of bytecode instructions; <tt class="xref py py-attr docutils literal">co_consts</tt> is a tuple containing the literals
used by the bytecode; <tt class="xref py py-attr docutils literal">co_names</tt> is a tuple containing the names used by
the bytecode; <tt class="xref py py-attr docutils literal">co_filename</tt> is the filename from which the code was
compiled; <tt class="xref py py-attr docutils literal">co_firstlineno</tt> is the first line number of the function;
<tt class="xref py py-attr docutils literal">co_lnotab</tt> is a string encoding the mapping from bytecode offsets to
line numbers (for details see the source code of the interpreter);
<tt class="xref py py-attr docutils literal">co_stacksize</tt> is the required stack size (including local variables);
<tt class="xref py py-attr docutils literal">co_flags</tt> is an integer encoding a number of flags for the interpreter.
The following flag bits are defined for <tt class="xref py py-attr docutils literal">co_flags</tt>: bit <tt class="docutils literal">0x04</tt> is set if
the function uses the <tt class="docutils literal">*arguments</tt> syntax to accept an arbitrary number of
positional arguments; bit <tt class="docutils literal">0x08</tt> is set if the function uses the
<tt class="docutils literal">**keywords</tt> syntax to accept arbitrary keyword arguments; bit <tt class="docutils literal">0x20</tt> is set
if the function is a generator.
Future feature declarations (<tt class="docutils literal">from __future__ import division</tt>) also use bits
in <tt class="xref py py-attr docutils literal">co_flags</tt> to indicate whether a code object was compiled with a
particular feature enabled: bit <tt class="docutils literal">0x2000</tt> is set if the function was compiled
with future division enabled; bits <tt class="docutils literal">0x10</tt> and <tt class="docutils literal">0x1000</tt> were used in earlier
versions of Python.
Other bits in <tt class="xref py py-attr docutils literal">co_flags</tt> are reserved for internal use.
If a code object represents a function, the first item in <tt class="xref py py-attr docutils literal">co_consts</tt> is
the documentation string of the function, or <tt class="docutils literal">None</tt> if undefined.
</dd>
</dl>
<dl class="last docutils" id="frame-objects">
<dt>Frame objects</dt>
<dd>Frame objects represent execution frames. They may occur in traceback objects
(see below).
Special read-only attributes: <tt class="xref py py-attr docutils literal">f_back</tt> is to the previous stack frame
(towards the caller), or <tt class="docutils literal">None</tt> if this is the bottom stack frame;
<tt class="xref py py-attr docutils literal">f_code</tt> is the code object being executed in this frame; <tt class="xref py py-attr docutils literal">f_locals</tt>
is the dictionary used to look up local variables; <tt class="xref py py-attr docutils literal">f_globals</tt> is used for
global variables; <tt class="xref py py-attr docutils literal">f_builtins</tt> is used for built-in (intrinsic) names;
<tt class="xref py py-attr docutils literal">f_restricted</tt> is a flag indicating whether the function is executing in
restricted execution mode; <tt class="xref py py-attr docutils literal">f_lasti</tt> gives the precise instruction (this
is an index into the bytecode string of the code object).
Special writable attributes: <tt class="xref py py-attr docutils literal">f_trace</tt>, if not <tt class="docutils literal">None</tt>, is a function
called at the start of each source code line (this is used by the debugger);
<tt class="xref py py-attr docutils literal">f_exc_type</tt>, <tt class="xref py py-attr docutils literal">f_exc_value</tt>, <tt class="xref py py-attr docutils literal">f_exc_traceback</tt> represent the
last exception raised in the parent frame provided another exception was ever
raised in the current frame (in all other cases they are None); <tt class="xref py py-attr docutils literal">f_lineno</tt>
is the current line number of the frame &#8212; writing to this from within a trace
function jumps to the given line (only for the bottom-most frame). A debugger
can implement a Jump command (aka Set Next Statement) by writing to f_lineno.
</dd>
<dt>Traceback objects</dt>
<dd>Traceback objects represent a stack trace of an exception. A traceback object
is created when an exception occurs. When the search for an exception handler
unwinds the execution stack, at each unwound level a traceback object is
inserted in front of the current traceback. When an exception handler is
entered, the stack trace is made available to the program. (See section
<a class="reference internal" href="compound_stmts.html#try">The try statement</a>.) It is accessible as <tt class="docutils literal">sys.exc_traceback</tt>,
and also as the third item of the tuple returned by <tt class="docutils literal">sys.exc_info()</tt>. The
latter is the preferred interface, since it works correctly when the program is
using multiple threads. When the program contains no suitable handler, the stack
trace is written (nicely formatted) to the standard error stream; if the
interpreter is interactive, it is also made available to the user as
<tt class="docutils literal">sys.last_traceback</tt>.
Special read-only attributes: <tt class="xref py py-attr docutils literal">tb_next</tt> is the next level in the stack
trace (towards the frame where the exception occurred), or <tt class="docutils literal">None</tt> if there is
no next level; <tt class="xref py py-attr docutils literal">tb_frame</tt> points to the execution frame of the current
level; <tt class="xref py py-attr docutils literal">tb_lineno</tt> gives the line number where the exception occurred;
<tt class="xref py py-attr docutils literal">tb_lasti</tt> indicates the precise instruction. The line number and last
instruction in the traceback may differ from the line number of its frame object
if the exception occurred in a <a class="reference internal" href="compound_stmts.html#try"><tt class="xref std std-keyword docutils literal">try</tt></a> statement with no matching except
clause or with a finally clause.
</dd>
<dt>Slice objects</dt>
<dd>Slice objects are used to represent slices when extended slice syntax is used.
This is a slice using two colons, or multiple slices or ellipses separated by
commas, e.g., <tt class="docutils literal">a[i:j:step]</tt>, <tt class="docutils literal">a[i:j, k:l]</tt>, or <tt class="docutils literal">a[..., i:j]</tt>. They are
also created by the built-in <a class="reference internal" href="../library/functions.html#slice" title="slice"><tt class="xref py py-func docutils literal">slice()</tt></a> function.
Special read-only attributes: <tt class="xref py py-attr docutils literal">start</tt> is the lower bound; <tt class="xref py py-attr docutils literal">stop</tt> is
the upper bound; <tt class="xref py py-attr docutils literal">step</tt> is the step value; each is <tt class="docutils literal">None</tt> if omitted.
These attributes can have any type.
Slice objects support one method:
<dl class="last method">
<dt id="slice.indices">
<tt class="descclassname">slice.</tt><tt class="descname">indices</tt><big>(</big>self, length<big>)</big><a class="headerlink" href="#slice.indices" title="Permalink to this definition">¶</a></dt>
<dd>This method takes a single integer argument length and computes information
about the extended slice that the slice object would describe if applied to a
sequence of length items. It returns a tuple of three integers; respectively
these are the start and stop indices and the step or stride length of the
slice. Missing or out-of-bounds indices are handled in a manner consistent with
regular slices.

New in version 2.3.
</dd></dl>

</dd>
<dt>Static method objects</dt>
<dd>Static method objects provide a way of defeating the transformation of function
objects to method objects described above. A static method object is a wrapper
around any other object, usually a user-defined method object. When a static
method object is retrieved from a class or a class instance, the object actually
returned is the wrapped object, which is not subject to any further
transformation. Static method objects are not themselves callable, although the
objects they wrap usually are. Static method objects are created by the built-in
<a class="reference internal" href="../library/functions.html#staticmethod" title="staticmethod"><tt class="xref py py-func docutils literal">staticmethod()</tt></a> constructor.</dd>
<dt>Class method objects</dt>
<dd>A class method object, like a static method object, is a wrapper around another
object that alters the way in which that object is retrieved from classes and
class instances. The behaviour of class method objects upon such retrieval is
described above, under &#8220;User-defined methods&#8221;. Class method objects are created
by the built-in <a class="reference internal" href="../library/functions.html#classmethod" title="classmethod"><tt class="xref py py-func docutils literal">classmethod()</tt></a> constructor.</dd>
</dl>
</dd>
</dl>
</div>
<div class="section" id="new-style-and-classic-classes">
<h2>3.3. New-style and classic classes<a class="headerlink" href="#new-style-and-classic-classes" title="Permalink to this headline">¶</a></h2>
Classes and instances come in two flavors: old-style (or classic) and new-style.
Up to Python 2.1, old-style classes were the only flavour available to the user.
The concept of (old-style) class is unrelated to the concept of type: if x is
an instance of an old-style class, then <tt class="docutils literal">x.__class__</tt> designates the class of
x, but <tt class="docutils literal">type(x)</tt> is always <tt class="docutils literal">&lt;type 'instance'&gt;</tt>. This reflects the fact
that all old-style instances, independently of their class, are implemented with
a single built-in type, called <tt class="docutils literal">instance</tt>.
New-style classes were introduced in Python 2.2 to unify classes and types. A
new-style class is neither more nor less than a user-defined type. If x is an
instance of a new-style class, then <tt class="docutils literal">type(x)</tt> is typically the same as
<tt class="docutils literal">x.__class__</tt> (although this is not guaranteed - a new-style class instance is
permitted to override the value returned for <tt class="docutils literal">x.__class__</tt>).
The major motivation for introducing new-style classes is to provide a unified
object model with a full meta-model. It also has a number of practical
benefits, like the ability to subclass most built-in types, or the introduction
of &#8220;descriptors&#8221;, which enable computed properties.
For compatibility reasons, classes are still old-style by default. New-style
classes are created by specifying another new-style class (i.e. a type) as a
parent class, or the &#8220;top-level type&#8221; <a class="reference internal" href="../library/functions.html#object" title="object"><tt class="xref py py-class docutils literal">object</tt></a> if no other parent is
needed. The behaviour of new-style classes differs from that of old-style
classes in a number of important details in addition to what <a class="reference internal" href="../library/functions.html#type" title="type"><tt class="xref py py-func docutils literal">type()</tt></a>
returns. Some of these changes are fundamental to the new object model, like
the way special methods are invoked. Others are &#8220;fixes&#8221; that could not be
implemented before for compatibility concerns, like the method resolution order
in case of multiple inheritance.
While this manual aims to provide comprehensive coverage of Python&#8217;s class
mechanics, it may still be lacking in some areas when it comes to its coverage
of new-style classes. Please see <a class="reference external" href="http://www.python.org/doc/newstyle/">http://www.python.org/doc/newstyle/</a> for
sources of additional information.
Old-style classes are removed in Python 3, leaving only the semantics of
new-style classes.
</div>
<div class="section" id="special-method-names">
<h2>3.4. Special method names<a class="headerlink" href="#special-method-names" title="Permalink to this headline">¶</a></h2>
A class can implement certain operations that are invoked by special syntax
(such as arithmetic operations or subscripting and slicing) by defining methods
with special names. This is Python&#8217;s approach to operator overloading,
allowing classes to define their own behavior with respect to language
operators. For instance, if a class defines a method named <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a>,
and <tt class="docutils literal">x</tt> is an instance of this class, then <tt class="docutils literal">x[i]</tt> is roughly equivalent
to <tt class="docutils literal">x.__getitem__(i)</tt> for old-style classes and <tt class="docutils literal">type(x).__getitem__(x, i)</tt>
for new-style classes. Except where mentioned, attempts to execute an
operation raise an exception when no appropriate method is defined (typically
<a class="reference internal" href="../library/exceptions.html#exceptions.AttributeError" title="exceptions.AttributeError"><tt class="xref py py-exc docutils literal">AttributeError</tt></a> or <a class="reference internal" href="../library/exceptions.html#exceptions.TypeError" title="exceptions.TypeError"><tt class="xref py py-exc docutils literal">TypeError</tt></a>).
When implementing a class that emulates any built-in type, it is important that
the emulation only be implemented to the degree that it makes sense for the
object being modelled. For example, some sequences may work well with retrieval
of individual elements, but extracting a slice may not make sense. (One example
of this is the <tt class="xref py py-class docutils literal">NodeList</tt> interface in the W3C&#8217;s Document Object Model.)
<div class="section" id="basic-customization">
<h3>3.4.1. Basic customization<a class="headerlink" href="#basic-customization" title="Permalink to this headline">¶</a></h3>
<dl class="method">
<dt id="object.__new__">
<tt class="descclassname">object.</tt><tt class="descname">__new__</tt><big>(</big>cls[, ...]<big>)</big><a class="headerlink" href="#object.__new__" title="Permalink to this definition">¶</a></dt>
<dd>Called to create a new instance of class cls. <a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a> is a static
method (special-cased so you need not declare it as such) that takes the class
of which an instance was requested as its first argument. The remaining
arguments are those passed to the object constructor expression (the call to the
class). The return value of <a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a> should be the new object instance
(usually an instance of cls).
Typical implementations create a new instance of the class by invoking the
superclass&#8217;s <a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a> method using <tt class="docutils literal">super(currentclass,
cls).__new__(cls[, ...])</tt> with appropriate arguments and then modifying the
newly-created instance as necessary before returning it.
If <a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a> returns an instance of cls, then the new instance&#8217;s
<a class="reference internal" href="#object.__init__" title="object.__init__"><tt class="xref py py-meth docutils literal">__init__()</tt></a> method will be invoked like <tt class="docutils literal">__init__(self[, ...])</tt>, where
self is the new instance and the remaining arguments are the same as were
passed to <a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a>.
If <a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a> does not return an instance of cls, then the new instance&#8217;s
<a class="reference internal" href="#object.__init__" title="object.__init__"><tt class="xref py py-meth docutils literal">__init__()</tt></a> method will not be invoked.
<a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a> is intended mainly to allow subclasses of immutable types (like
int, str, or tuple) to customize instance creation. It is also commonly
overridden in custom metaclasses in order to customize class creation.
</dd></dl>

<dl class="method">
<dt id="object.__init__">
<tt class="descclassname">object.</tt><tt class="descname">__init__</tt><big>(</big>self[, ...]<big>)</big><a class="headerlink" href="#object.__init__" title="Permalink to this definition">¶</a></dt>
<dd>Called when the instance is created. The arguments are those passed to the
class constructor expression. If a base class has an <a class="reference internal" href="#object.__init__" title="object.__init__"><tt class="xref py py-meth docutils literal">__init__()</tt></a> method,
the derived class&#8217;s <a class="reference internal" href="#object.__init__" title="object.__init__"><tt class="xref py py-meth docutils literal">__init__()</tt></a> method, if any, must explicitly call it to
ensure proper initialization of the base class part of the instance; for
example: <tt class="docutils literal">BaseClass.__init__(self, [args...])</tt>. As a special constraint on
constructors, no value may be returned; doing so will cause a <a class="reference internal" href="../library/exceptions.html#exceptions.TypeError" title="exceptions.TypeError"><tt class="xref py py-exc docutils literal">TypeError</tt></a>
to be raised at runtime.
</dd></dl>

<dl class="method">
<dt id="object.__del__">
<tt class="descclassname">object.</tt><tt class="descname">__del__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__del__" title="Permalink to this definition">¶</a></dt>
<dd>Called when the instance is about to be destroyed. This is also called a
destructor. If a base class has a <a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> method, the derived class&#8217;s
<a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> method, if any, must explicitly call it to ensure proper
deletion of the base class part of the instance. Note that it is possible
(though not recommended!) for the <a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> method to postpone destruction
of the instance by creating a new reference to it. It may then be called at a
later time when this new reference is deleted. It is not guaranteed that
<a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> methods are called for objects that still exist when the
interpreter exits.
<div class="admonition note">
Note
<tt class="docutils literal">del x</tt> doesn&#8217;t directly call <tt class="docutils literal">x.__del__()</tt> &#8212; the former decrements
the reference count for <tt class="docutils literal">x</tt> by one, and the latter is only called when
<tt class="docutils literal">x</tt>&#8216;s reference count reaches zero. Some common situations that may
prevent the reference count of an object from going to zero include:
circular references between objects (e.g., a doubly-linked list or a tree
data structure with parent and child pointers); a reference to the object
on the stack frame of a function that caught an exception (the traceback
stored in <tt class="docutils literal">sys.exc_traceback</tt> keeps the stack frame alive); or a
reference to the object on the stack frame that raised an unhandled
exception in interactive mode (the traceback stored in
<tt class="docutils literal">sys.last_traceback</tt> keeps the stack frame alive). The first situation
can only be remedied by explicitly breaking the cycles; the latter two
situations can be resolved by storing <tt class="docutils literal">None</tt> in <tt class="docutils literal">sys.exc_traceback</tt> or
<tt class="docutils literal">sys.last_traceback</tt>. Circular references which are garbage are
detected when the option cycle detector is enabled (it&#8217;s on by default),
but can only be cleaned up if there are no Python-level <a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a>
methods involved. Refer to the documentation for the <a class="reference internal" href="../library/gc.html#module-gc" title="gc: Interface to the cycle-detecting garbage collector."><tt class="xref py py-mod docutils literal">gc</tt></a> module for
more information about how <a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> methods are handled by the
cycle detector, particularly the description of the <tt class="docutils literal">garbage</tt> value.
</div>
<div class="admonition warning">
Warning
Due to the precarious circumstances under which <a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> methods are
invoked, exceptions that occur during their execution are ignored, and a warning
is printed to <tt class="docutils literal">sys.stderr</tt> instead. Also, when <a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> is invoked in
response to a module being deleted (e.g., when execution of the program is
done), other globals referenced by the <a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> method may already have
been deleted or in the process of being torn down (e.g. the import
machinery shutting down). For this reason, <a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> methods
should do the absolute
minimum needed to maintain external invariants. Starting with version 1.5,
Python guarantees that globals whose name begins with a single underscore are
deleted from their module before other globals are deleted; if no other
references to such globals exist, this may help in assuring that imported
modules are still available at the time when the <a class="reference internal" href="#object.__del__" title="object.__del__"><tt class="xref py py-meth docutils literal">__del__()</tt></a> method is
called.
</div>
See also the <a class="reference internal" href="../using/cmdline.html#cmdoption-R">-R</a> command-line option.
</dd></dl>

<dl class="method">
<dt id="object.__repr__">
<tt class="descclassname">object.</tt><tt class="descname">__repr__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__repr__" title="Permalink to this definition">¶</a></dt>
<dd>Called by the <a class="reference internal" href="../library/repr.html#module-repr" title="repr: Alternate repr() implementation with size limits."><tt class="xref py py-func docutils literal">repr()</tt></a> built-in function and by string conversions (reverse
quotes) to compute the &#8220;official&#8221; string representation of an object. If at all
possible, this should look like a valid Python expression that could be used to
recreate an object with the same value (given an appropriate environment). If
this is not possible, a string of the form <tt class="docutils literal">&lt;...some useful description...&gt;</tt>
should be returned. The return value must be a string object. If a class
defines <a class="reference internal" href="#object.__repr__" title="object.__repr__"><tt class="xref py py-meth docutils literal">__repr__()</tt></a> but not <a class="reference internal" href="#object.__str__" title="object.__str__"><tt class="xref py py-meth docutils literal">__str__()</tt></a>, then <a class="reference internal" href="#object.__repr__" title="object.__repr__"><tt class="xref py py-meth docutils literal">__repr__()</tt></a> is also
used when an &#8220;informal&#8221; string representation of instances of that class is
required.
This is typically used for debugging, so it is important that the representation
is information-rich and unambiguous.
</dd></dl>

<dl class="method">
<dt id="object.__str__">
<tt class="descclassname">object.</tt><tt class="descname">__str__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__str__" title="Permalink to this definition">¶</a></dt>
<dd>Called by the <a class="reference internal" href="../library/functions.html#str" title="str"><tt class="xref py py-func docutils literal">str()</tt></a> built-in function and by the <a class="reference internal" href="simple_stmts.html#print"><tt class="xref std std-keyword docutils literal">print</tt></a>
statement to compute the &#8220;informal&#8221; string representation of an object. This
differs from <a class="reference internal" href="#object.__repr__" title="object.__repr__"><tt class="xref py py-meth docutils literal">__repr__()</tt></a> in that it does not have to be a valid Python
expression: a more convenient or concise representation may be used instead.
The return value must be a string object.
</dd></dl>

<dl class="method">
<dt id="object.__lt__">
<tt class="descclassname">object.</tt><tt class="descname">__lt__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__lt__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__le__">
<tt class="descclassname">object.</tt><tt class="descname">__le__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__le__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__eq__">
<tt class="descclassname">object.</tt><tt class="descname">__eq__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__eq__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__ne__">
<tt class="descclassname">object.</tt><tt class="descname">__ne__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__ne__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__gt__">
<tt class="descclassname">object.</tt><tt class="descname">__gt__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__gt__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__ge__">
<tt class="descclassname">object.</tt><tt class="descname">__ge__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__ge__" title="Permalink to this definition">¶</a></dt>
<dd>
New in version 2.1.
These are the so-called &#8220;rich comparison&#8221; methods, and are called for comparison
operators in preference to <a class="reference internal" href="#object.__cmp__" title="object.__cmp__"><tt class="xref py py-meth docutils literal">__cmp__()</tt></a> below. The correspondence between
operator symbols and method names is as follows: <tt class="docutils literal">x&lt;y</tt> calls <tt class="docutils literal">x.__lt__(y)</tt>,
<tt class="docutils literal">x&lt;=y</tt> calls <tt class="docutils literal">x.__le__(y)</tt>, <tt class="docutils literal">x==y</tt> calls <tt class="docutils literal">x.__eq__(y)</tt>, <tt class="docutils literal">x!=y</tt> and
<tt class="docutils literal">x&lt;&gt;y</tt> call <tt class="docutils literal">x.__ne__(y)</tt>, <tt class="docutils literal">x&gt;y</tt> calls <tt class="docutils literal">x.__gt__(y)</tt>, and <tt class="docutils literal">x&gt;=y</tt> calls
<tt class="docutils literal">x.__ge__(y)</tt>.
A rich comparison method may return the singleton <tt class="docutils literal">NotImplemented</tt> if it does
not implement the operation for a given pair of arguments. By convention,
<tt class="docutils literal">False</tt> and <tt class="docutils literal">True</tt> are returned for a successful comparison. However, these
methods can return any value, so if the comparison operator is used in a Boolean
context (e.g., in the condition of an <tt class="docutils literal">if</tt> statement), Python will call
<a class="reference internal" href="../library/functions.html#bool" title="bool"><tt class="xref py py-func docutils literal">bool()</tt></a> on the value to determine if the result is true or false.
There are no implied relationships among the comparison operators. The truth
of <tt class="docutils literal">x==y</tt> does not imply that <tt class="docutils literal">x!=y</tt> is false. Accordingly, when
defining <a class="reference internal" href="#object.__eq__" title="object.__eq__"><tt class="xref py py-meth docutils literal">__eq__()</tt></a>, one should also define <a class="reference internal" href="#object.__ne__" title="object.__ne__"><tt class="xref py py-meth docutils literal">__ne__()</tt></a> so that the
operators will behave as expected. See the paragraph on <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a> for
some important notes on creating <a class="reference internal" href="../glossary.html#term-hashable">hashable</a> objects which support
custom comparison operations and are usable as dictionary keys.
There are no swapped-argument versions of these methods (to be used when the
left argument does not support the operation but the right argument does);
rather, <a class="reference internal" href="#object.__lt__" title="object.__lt__"><tt class="xref py py-meth docutils literal">__lt__()</tt></a> and <a class="reference internal" href="#object.__gt__" title="object.__gt__"><tt class="xref py py-meth docutils literal">__gt__()</tt></a> are each other&#8217;s reflection,
<a class="reference internal" href="#object.__le__" title="object.__le__"><tt class="xref py py-meth docutils literal">__le__()</tt></a> and <a class="reference internal" href="#object.__ge__" title="object.__ge__"><tt class="xref py py-meth docutils literal">__ge__()</tt></a> are each other&#8217;s reflection, and
<a class="reference internal" href="#object.__eq__" title="object.__eq__"><tt class="xref py py-meth docutils literal">__eq__()</tt></a> and <a class="reference internal" href="#object.__ne__" title="object.__ne__"><tt class="xref py py-meth docutils literal">__ne__()</tt></a> are their own reflection.
Arguments to rich comparison methods are never coerced.
To automatically generate ordering operations from a single root operation,
see <a class="reference internal" href="../library/functools.html#functools.total_ordering" title="functools.total_ordering"><tt class="xref py py-func docutils literal">functools.total_ordering()</tt></a>.
</dd></dl>

<dl class="method">
<dt id="object.__cmp__">
<tt class="descclassname">object.</tt><tt class="descname">__cmp__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__cmp__" title="Permalink to this definition">¶</a></dt>
<dd>Called by comparison operations if rich comparison (see above) is not
defined. Should return a negative integer if <tt class="docutils literal">self &lt; other</tt>, zero if
<tt class="docutils literal">self == other</tt>, a positive integer if <tt class="docutils literal">self &gt; other</tt>. If no
<a class="reference internal" href="#object.__cmp__" title="object.__cmp__"><tt class="xref py py-meth docutils literal">__cmp__()</tt></a>, <a class="reference internal" href="#object.__eq__" title="object.__eq__"><tt class="xref py py-meth docutils literal">__eq__()</tt></a> or <a class="reference internal" href="#object.__ne__" title="object.__ne__"><tt class="xref py py-meth docutils literal">__ne__()</tt></a> operation is defined, class
instances are compared by object identity (&#8220;address&#8221;). See also the
description of <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a> for some important notes on creating
<a class="reference internal" href="../glossary.html#term-hashable">hashable</a> objects which support custom comparison operations and are
usable as dictionary keys. (Note: the restriction that exceptions are not
propagated by <a class="reference internal" href="#object.__cmp__" title="object.__cmp__"><tt class="xref py py-meth docutils literal">__cmp__()</tt></a> has been removed since Python 1.5.)
</dd></dl>

<dl class="method">
<dt id="object.__rcmp__">
<tt class="descclassname">object.</tt><tt class="descname">__rcmp__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rcmp__" title="Permalink to this definition">¶</a></dt>
<dd>
Changed in version 2.1: No longer supported.
</dd></dl>

<dl class="method">
<dt id="object.__hash__">
<tt class="descclassname">object.</tt><tt class="descname">__hash__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__hash__" title="Permalink to this definition">¶</a></dt>
<dd>Called by built-in function <a class="reference internal" href="../library/functions.html#hash" title="hash"><tt class="xref py py-func docutils literal">hash()</tt></a> and for operations on members of
hashed collections including <a class="reference internal" href="../library/stdtypes.html#set" title="set"><tt class="xref py py-class docutils literal">set</tt></a>, <a class="reference internal" href="../library/stdtypes.html#frozenset" title="frozenset"><tt class="xref py py-class docutils literal">frozenset</tt></a>, and
<a class="reference internal" href="../library/stdtypes.html#dict" title="dict"><tt class="xref py py-class docutils literal">dict</tt></a>. <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a> should return an integer. The only required
property is that objects which compare equal have the same hash value; it is
advised to somehow mix together (e.g. using exclusive or) the hash values for
the components of the object that also play a part in comparison of objects.
If a class does not define a <a class="reference internal" href="#object.__cmp__" title="object.__cmp__"><tt class="xref py py-meth docutils literal">__cmp__()</tt></a> or <a class="reference internal" href="#object.__eq__" title="object.__eq__"><tt class="xref py py-meth docutils literal">__eq__()</tt></a> method it
should not define a <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a> operation either; if it defines
<a class="reference internal" href="#object.__cmp__" title="object.__cmp__"><tt class="xref py py-meth docutils literal">__cmp__()</tt></a> or <a class="reference internal" href="#object.__eq__" title="object.__eq__"><tt class="xref py py-meth docutils literal">__eq__()</tt></a> but not <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a>, its instances
will not be usable in hashed collections. If a class defines mutable objects
and implements a <a class="reference internal" href="#object.__cmp__" title="object.__cmp__"><tt class="xref py py-meth docutils literal">__cmp__()</tt></a> or <a class="reference internal" href="#object.__eq__" title="object.__eq__"><tt class="xref py py-meth docutils literal">__eq__()</tt></a> method, it should not
implement <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a>, since hashable collection implementations require
that a object&#8217;s hash value is immutable (if the object&#8217;s hash value changes,
it will be in the wrong hash bucket).
User-defined classes have <a class="reference internal" href="#object.__cmp__" title="object.__cmp__"><tt class="xref py py-meth docutils literal">__cmp__()</tt></a> and <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a> methods
by default; with them, all objects compare unequal (except with themselves)
and <tt class="docutils literal">x.__hash__()</tt> returns <tt class="docutils literal">id(x)</tt>.
Classes which inherit a <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a> method from a parent class but
change the meaning of <a class="reference internal" href="#object.__cmp__" title="object.__cmp__"><tt class="xref py py-meth docutils literal">__cmp__()</tt></a> or <a class="reference internal" href="#object.__eq__" title="object.__eq__"><tt class="xref py py-meth docutils literal">__eq__()</tt></a> such that the hash
value returned is no longer appropriate (e.g. by switching to a value-based
concept of equality instead of the default identity based equality) can
explicitly flag themselves as being unhashable by setting <tt class="docutils literal">__hash__ = None</tt>
in the class definition. Doing so means that not only will instances of the
class raise an appropriate <a class="reference internal" href="../library/exceptions.html#exceptions.TypeError" title="exceptions.TypeError"><tt class="xref py py-exc docutils literal">TypeError</tt></a> when a program attempts to
retrieve their hash value, but they will also be correctly identified as
unhashable when checking <tt class="docutils literal">isinstance(obj, collections.Hashable)</tt> (unlike
classes which define their own <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a> to explicitly raise
<a class="reference internal" href="../library/exceptions.html#exceptions.TypeError" title="exceptions.TypeError"><tt class="xref py py-exc docutils literal">TypeError</tt></a>).

Changed in version 2.5: <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a> may now also return a long integer object; the 32-bit
integer is then derived from the hash of that object.

Changed in version 2.6: <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-attr docutils literal">__hash__</tt></a> may now be set to <a class="reference internal" href="../library/constants.html#None" title="None"><tt class="xref py py-const docutils literal">None</tt></a> to explicitly flag
instances of a class as unhashable.
</dd></dl>

<dl class="method">
<dt id="object.__nonzero__">
<tt class="descclassname">object.</tt><tt class="descname">__nonzero__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__nonzero__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement truth value testing and the built-in operation <tt class="docutils literal">bool()</tt>;
should return <tt class="docutils literal">False</tt> or <tt class="docutils literal">True</tt>, or their integer equivalents <tt class="docutils literal">0</tt> or
<tt class="docutils literal">1</tt>. When this method is not defined, <a class="reference internal" href="#object.__len__" title="object.__len__"><tt class="xref py py-meth docutils literal">__len__()</tt></a> is called, if it is
defined, and the object is considered true if its result is nonzero.
If a class defines neither <a class="reference internal" href="#object.__len__" title="object.__len__"><tt class="xref py py-meth docutils literal">__len__()</tt></a> nor <a class="reference internal" href="#object.__nonzero__" title="object.__nonzero__"><tt class="xref py py-meth docutils literal">__nonzero__()</tt></a>, all its
instances are considered true.
</dd></dl>

<dl class="method">
<dt id="object.__unicode__">
<tt class="descclassname">object.</tt><tt class="descname">__unicode__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__unicode__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement <a class="reference internal" href="../library/functions.html#unicode" title="unicode"><tt class="xref py py-func docutils literal">unicode()</tt></a> built-in; should return a Unicode object.
When this method is not defined, string conversion is attempted, and the result
of string conversion is converted to Unicode using the system default encoding.
</dd></dl>

</div>
<div class="section" id="customizing-attribute-access">
<h3>3.4.2. Customizing attribute access<a class="headerlink" href="#customizing-attribute-access" title="Permalink to this headline">¶</a></h3>
The following methods can be defined to customize the meaning of attribute
access (use of, assignment to, or deletion of <tt class="docutils literal">x.name</tt>) for class instances.
<dl class="method">
<dt id="object.__getattr__">
<tt class="descclassname">object.</tt><tt class="descname">__getattr__</tt><big>(</big>self, name<big>)</big><a class="headerlink" href="#object.__getattr__" title="Permalink to this definition">¶</a></dt>
<dd>Called when an attribute lookup has not found the attribute in the usual places
(i.e. it is not an instance attribute nor is it found in the class tree for
<tt class="docutils literal">self</tt>). <tt class="docutils literal">name</tt> is the attribute name. This method should return the
(computed) attribute value or raise an <a class="reference internal" href="../library/exceptions.html#exceptions.AttributeError" title="exceptions.AttributeError"><tt class="xref py py-exc docutils literal">AttributeError</tt></a> exception.
Note that if the attribute is found through the normal mechanism,
<a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><tt class="xref py py-meth docutils literal">__getattr__()</tt></a> is not called. (This is an intentional asymmetry between
<a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><tt class="xref py py-meth docutils literal">__getattr__()</tt></a> and <a class="reference internal" href="#object.__setattr__" title="object.__setattr__"><tt class="xref py py-meth docutils literal">__setattr__()</tt></a>.) This is done both for efficiency
reasons and because otherwise <a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><tt class="xref py py-meth docutils literal">__getattr__()</tt></a> would have no way to access
other attributes of the instance. Note that at least for instance variables,
you can fake total control by not inserting any values in the instance attribute
dictionary (but instead inserting them in another object). See the
<a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><tt class="xref py py-meth docutils literal">__getattribute__()</tt></a> method below for a way to actually get total control in
new-style classes.
</dd></dl>

<dl class="method">
<dt id="object.__setattr__">
<tt class="descclassname">object.</tt><tt class="descname">__setattr__</tt><big>(</big>self, name, value<big>)</big><a class="headerlink" href="#object.__setattr__" title="Permalink to this definition">¶</a></dt>
<dd>Called when an attribute assignment is attempted. This is called instead of the
normal mechanism (i.e. store the value in the instance dictionary). name is
the attribute name, value is the value to be assigned to it.
If <a class="reference internal" href="#object.__setattr__" title="object.__setattr__"><tt class="xref py py-meth docutils literal">__setattr__()</tt></a> wants to assign to an instance attribute, it should not
simply execute <tt class="docutils literal">self.name = value</tt> &#8212; this would cause a recursive call to
itself. Instead, it should insert the value in the dictionary of instance
attributes, e.g., <tt class="docutils literal">self.__dict__[name] = value</tt>. For new-style classes,
rather than accessing the instance dictionary, it should call the base class
method with the same name, for example, <tt class="docutils literal">object.__setattr__(self, name,
value)</tt>.
</dd></dl>

<dl class="method">
<dt id="object.__delattr__">
<tt class="descclassname">object.</tt><tt class="descname">__delattr__</tt><big>(</big>self, name<big>)</big><a class="headerlink" href="#object.__delattr__" title="Permalink to this definition">¶</a></dt>
<dd>Like <a class="reference internal" href="#object.__setattr__" title="object.__setattr__"><tt class="xref py py-meth docutils literal">__setattr__()</tt></a> but for attribute deletion instead of assignment. This
should only be implemented if <tt class="docutils literal">del obj.name</tt> is meaningful for the object.
</dd></dl>

<div class="section" id="more-attribute-access-for-new-style-classes">
<h4>3.4.2.1. More attribute access for new-style classes<a class="headerlink" href="#more-attribute-access-for-new-style-classes" title="Permalink to this headline">¶</a></h4>
The following methods only apply to new-style classes.
<dl class="method">
<dt id="object.__getattribute__">
<tt class="descclassname">object.</tt><tt class="descname">__getattribute__</tt><big>(</big>self, name<big>)</big><a class="headerlink" href="#object.__getattribute__" title="Permalink to this definition">¶</a></dt>
<dd>Called unconditionally to implement attribute accesses for instances of the
class. If the class also defines <a class="reference internal" href="#object.__getattr__" title="object.__getattr__"><tt class="xref py py-meth docutils literal">__getattr__()</tt></a>, the latter will not be
called unless <a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><tt class="xref py py-meth docutils literal">__getattribute__()</tt></a> either calls it explicitly or raises an
<a class="reference internal" href="../library/exceptions.html#exceptions.AttributeError" title="exceptions.AttributeError"><tt class="xref py py-exc docutils literal">AttributeError</tt></a>. This method should return the (computed) attribute value
or raise an <a class="reference internal" href="../library/exceptions.html#exceptions.AttributeError" title="exceptions.AttributeError"><tt class="xref py py-exc docutils literal">AttributeError</tt></a> exception. In order to avoid infinite
recursion in this method, its implementation should always call the base class
method with the same name to access any attributes it needs, for example,
<tt class="docutils literal">object.__getattribute__(self, name)</tt>.
<div class="admonition note">
Note
This method may still be bypassed when looking up special methods as the
result of implicit invocation via language syntax or built-in functions.
See <a class="reference internal" href="#new-style-special-lookup">Special method lookup for new-style classes</a>.
</div>
</dd></dl>

</div>
<div class="section" id="implementing-descriptors">
<h4>3.4.2.2. Implementing Descriptors<a class="headerlink" href="#implementing-descriptors" title="Permalink to this headline">¶</a></h4>
The following methods only apply when an instance of the class containing the
method (a so-called descriptor class) appears in an owner class (the
descriptor must be in either the owner&#8217;s class dictionary or in the class
dictionary for one of its parents). In the examples below, &#8220;the attribute&#8221;
refers to the attribute whose name is the key of the property in the owner
class&#8217; <tt class="xref py py-attr docutils literal">__dict__</tt>.
<dl class="method">
<dt id="object.__get__">
<tt class="descclassname">object.</tt><tt class="descname">__get__</tt><big>(</big>self, instance, owner<big>)</big><a class="headerlink" href="#object.__get__" title="Permalink to this definition">¶</a></dt>
<dd>Called to get the attribute of the owner class (class attribute access) or of an
instance of that class (instance attribute access). owner is always the owner
class, while instance is the instance that the attribute was accessed through,
or <tt class="docutils literal">None</tt> when the attribute is accessed through the owner. This method
should return the (computed) attribute value or raise an <a class="reference internal" href="../library/exceptions.html#exceptions.AttributeError" title="exceptions.AttributeError"><tt class="xref py py-exc docutils literal">AttributeError</tt></a>
exception.
</dd></dl>

<dl class="method">
<dt id="object.__set__">
<tt class="descclassname">object.</tt><tt class="descname">__set__</tt><big>(</big>self, instance, value<big>)</big><a class="headerlink" href="#object.__set__" title="Permalink to this definition">¶</a></dt>
<dd>Called to set the attribute on an instance instance of the owner class to a
new value, value.
</dd></dl>

<dl class="method">
<dt id="object.__delete__">
<tt class="descclassname">object.</tt><tt class="descname">__delete__</tt><big>(</big>self, instance<big>)</big><a class="headerlink" href="#object.__delete__" title="Permalink to this definition">¶</a></dt>
<dd>Called to delete the attribute on an instance instance of the owner class.
</dd></dl>

</div>
<div class="section" id="invoking-descriptors">
<h4>3.4.2.3. Invoking Descriptors<a class="headerlink" href="#invoking-descriptors" title="Permalink to this headline">¶</a></h4>
In general, a descriptor is an object attribute with &#8220;binding behavior&#8221;, one
whose attribute access has been overridden by methods in the descriptor
protocol: <a class="reference internal" href="#object.__get__" title="object.__get__"><tt class="xref py py-meth docutils literal">__get__()</tt></a>, <a class="reference internal" href="#object.__set__" title="object.__set__"><tt class="xref py py-meth docutils literal">__set__()</tt></a>, and <a class="reference internal" href="#object.__delete__" title="object.__delete__"><tt class="xref py py-meth docutils literal">__delete__()</tt></a>. If any of
those methods are defined for an object, it is said to be a descriptor.
The default behavior for attribute access is to get, set, or delete the
attribute from an object&#8217;s dictionary. For instance, <tt class="docutils literal">a.x</tt> has a lookup chain
starting with <tt class="docutils literal">a.__dict__['x']</tt>, then <tt class="docutils literal">type(a).__dict__['x']</tt>, and
continuing through the base classes of <tt class="docutils literal">type(a)</tt> excluding metaclasses.
However, if the looked-up value is an object defining one of the descriptor
methods, then Python may override the default behavior and invoke the descriptor
method instead. Where this occurs in the precedence chain depends on which
descriptor methods were defined and how they were called. Note that descriptors
are only invoked for new style objects or classes (ones that subclass
<a class="reference internal" href="../library/functions.html#object" title="object"><tt class="xref py py-class docutils literal">object()</tt></a> or <a class="reference internal" href="../library/functions.html#type" title="type"><tt class="xref py py-class docutils literal">type()</tt></a>).
The starting point for descriptor invocation is a binding, <tt class="docutils literal">a.x</tt>. How the
arguments are assembled depends on <tt class="docutils literal">a</tt>:
<dl class="docutils">
<dt>Direct Call</dt>
<dd>The simplest and least common call is when user code directly invokes a
descriptor method: <tt class="docutils literal">x.__get__(a)</tt>.</dd>
<dt>Instance Binding</dt>
<dd>If binding to a new-style object instance, <tt class="docutils literal">a.x</tt> is transformed into the call:
<tt class="docutils literal">type(a).__dict__['x'].__get__(a, type(a))</tt>.</dd>
<dt>Class Binding</dt>
<dd>If binding to a new-style class, <tt class="docutils literal">A.x</tt> is transformed into the call:
<tt class="docutils literal">A.__dict__['x'].__get__(None, A)</tt>.</dd>
<dt>Super Binding</dt>
<dd>If <tt class="docutils literal">a</tt> is an instance of <a class="reference internal" href="../library/functions.html#super" title="super"><tt class="xref py py-class docutils literal">super</tt></a>, then the binding <tt class="docutils literal">super(B,
obj).m()</tt> searches <tt class="docutils literal">obj.__class__.__mro__</tt> for the base class <tt class="docutils literal">A</tt>
immediately preceding <tt class="docutils literal">B</tt> and then invokes the descriptor with the call:
<tt class="docutils literal">A.__dict__['m'].__get__(obj, obj.__class__)</tt>.</dd>
</dl>
For instance bindings, the precedence of descriptor invocation depends on the
which descriptor methods are defined. A descriptor can define any combination
of <a class="reference internal" href="#object.__get__" title="object.__get__"><tt class="xref py py-meth docutils literal">__get__()</tt></a>, <a class="reference internal" href="#object.__set__" title="object.__set__"><tt class="xref py py-meth docutils literal">__set__()</tt></a> and <a class="reference internal" href="#object.__delete__" title="object.__delete__"><tt class="xref py py-meth docutils literal">__delete__()</tt></a>. If it does not
define <a class="reference internal" href="#object.__get__" title="object.__get__"><tt class="xref py py-meth docutils literal">__get__()</tt></a>, then accessing the attribute will return the descriptor
object itself unless there is a value in the object&#8217;s instance dictionary. If
the descriptor defines <a class="reference internal" href="#object.__set__" title="object.__set__"><tt class="xref py py-meth docutils literal">__set__()</tt></a> and/or <a class="reference internal" href="#object.__delete__" title="object.__delete__"><tt class="xref py py-meth docutils literal">__delete__()</tt></a>, it is a data
descriptor; if it defines neither, it is a non-data descriptor. Normally, data
descriptors define both <a class="reference internal" href="#object.__get__" title="object.__get__"><tt class="xref py py-meth docutils literal">__get__()</tt></a> and <a class="reference internal" href="#object.__set__" title="object.__set__"><tt class="xref py py-meth docutils literal">__set__()</tt></a>, while non-data
descriptors have just the <a class="reference internal" href="#object.__get__" title="object.__get__"><tt class="xref py py-meth docutils literal">__get__()</tt></a> method. Data descriptors with
<a class="reference internal" href="#object.__set__" title="object.__set__"><tt class="xref py py-meth docutils literal">__set__()</tt></a> and <a class="reference internal" href="#object.__get__" title="object.__get__"><tt class="xref py py-meth docutils literal">__get__()</tt></a> defined always override a redefinition in an
instance dictionary. In contrast, non-data descriptors can be overridden by
instances.
Python methods (including <a class="reference internal" href="../library/functions.html#staticmethod" title="staticmethod"><tt class="xref py py-func docutils literal">staticmethod()</tt></a> and <a class="reference internal" href="../library/functions.html#classmethod" title="classmethod"><tt class="xref py py-func docutils literal">classmethod()</tt></a>) are
implemented as non-data descriptors. Accordingly, instances can redefine and
override methods. This allows individual instances to acquire behaviors that
differ from other instances of the same class.
The <a class="reference internal" href="../library/functions.html#property" title="property"><tt class="xref py py-func docutils literal">property()</tt></a> function is implemented as a data descriptor. Accordingly,
instances cannot override the behavior of a property.
</div>
<div class="section" id="slots">
<h4>3.4.2.4. __slots__<a class="headerlink" href="#slots" title="Permalink to this headline">¶</a></h4>
By default, instances of both old and new-style classes have a dictionary for
attribute storage. This wastes space for objects having very few instance
variables. The space consumption can become acute when creating large numbers
of instances.
The default can be overridden by defining __slots__ in a new-style class
definition. The __slots__ declaration takes a sequence of instance variables
and reserves just enough space in each instance to hold a value for each
variable. Space is saved because __dict__ is not created for each instance.
<dl class="data">
<dt id="__slots__">
<tt class="descname">__slots__</tt><a class="headerlink" href="#__slots__" title="Permalink to this definition">¶</a></dt>
<dd>This class variable can be assigned a string, iterable, or sequence of strings
with variable names used by instances. If defined in a new-style class,
__slots__ reserves space for the declared variables and prevents the automatic
creation of __dict__ and __weakref__ for each instance.

New in version 2.2.
</dd></dl>

Notes on using __slots__
<ul>
<li>When inheriting from a class without __slots__, the __dict__ attribute of
that class will always be accessible, so a __slots__ definition in the
subclass is meaningless.
</li>
<li>Without a __dict__ variable, instances cannot be assigned new variables not
listed in the __slots__ definition. Attempts to assign to an unlisted
variable name raises <a class="reference internal" href="../library/exceptions.html#exceptions.AttributeError" title="exceptions.AttributeError"><tt class="xref py py-exc docutils literal">AttributeError</tt></a>. If dynamic assignment of new
variables is desired, then add <tt class="docutils literal">'__dict__'</tt> to the sequence of strings in the
__slots__ declaration.

Changed in version 2.3: Previously, adding <tt class="docutils literal">'__dict__'</tt> to the __slots__ declaration would not
enable the assignment of new attributes not specifically listed in the sequence
of instance variable names.
</li>
<li>Without a __weakref__ variable for each instance, classes defining
__slots__ do not support weak references to its instances. If weak reference
support is needed, then add <tt class="docutils literal">'__weakref__'</tt> to the sequence of strings in the
__slots__ declaration.

Changed in version 2.3: Previously, adding <tt class="docutils literal">'__weakref__'</tt> to the __slots__ declaration would not
enable support for weak references.
</li>
<li>__slots__ are implemented at the class level by creating descriptors
(<a class="reference internal" href="#descriptors">Implementing Descriptors</a>) for each variable name. As a result, class attributes
cannot be used to set default values for instance variables defined by
__slots__; otherwise, the class attribute would overwrite the descriptor
assignment.
</li>
<li>The action of a __slots__ declaration is limited to the class where it is
defined. As a result, subclasses will have a __dict__ unless they also define
__slots__ (which must only contain names of any additional slots).
</li>
<li>If a class defines a slot also defined in a base class, the instance variable
defined by the base class slot is inaccessible (except by retrieving its
descriptor directly from the base class). This renders the meaning of the
program undefined. In the future, a check may be added to prevent this.
</li>
<li>Nonempty __slots__ does not work for classes derived from &#8220;variable-length&#8221;
built-in types such as <a class="reference internal" href="../library/functions.html#long" title="long"><tt class="xref py py-class docutils literal">long</tt></a>, <a class="reference internal" href="../library/functions.html#str" title="str"><tt class="xref py py-class docutils literal">str</tt></a> and <a class="reference internal" href="../library/functions.html#tuple" title="tuple"><tt class="xref py py-class docutils literal">tuple</tt></a>.
</li>
<li>Any non-string iterable may be assigned to __slots__. Mappings may also be
used; however, in the future, special meaning may be assigned to the values
corresponding to each key.
</li>
<li>__class__ assignment works only if both classes have the same __slots__.

Changed in version 2.6: Previously, __class__ assignment raised an error if either new or old class
had __slots__.
</li>
</ul>
</div>
</div>
<div class="section" id="customizing-class-creation">
<h3>3.4.3. Customizing class creation<a class="headerlink" href="#customizing-class-creation" title="Permalink to this headline">¶</a></h3>
By default, new-style classes are constructed using <a class="reference internal" href="../library/functions.html#type" title="type"><tt class="xref py py-func docutils literal">type()</tt></a>. A class
definition is read into a separate namespace and the value of class name is
bound to the result of <tt class="docutils literal">type(name, bases, dict)</tt>.
When the class definition is read, if __metaclass__ is defined then the
callable assigned to it will be called instead of <a class="reference internal" href="../library/functions.html#type" title="type"><tt class="xref py py-func docutils literal">type()</tt></a>. This allows
classes or functions to be written which monitor or alter the class creation
process:
<ul class="simple">
<li>Modifying the class dictionary prior to the class being created.</li>
<li>Returning an instance of another class &#8211; essentially performing the role of a
factory function.</li>
</ul>
These steps will have to be performed in the metaclass&#8217;s <a class="reference internal" href="#object.__new__" title="object.__new__"><tt class="xref py py-meth docutils literal">__new__()</tt></a> method
&#8211; <tt class="xref py py-meth docutils literal">type.__new__()</tt> can then be called from this method to create a class
with different properties. This example adds a new element to the class
dictionary before creating the class:
<div class="highlight-python"><div class="highlight"><pre>class metacls(type):
 def __new__(mcs, name, bases, dict):
 dict[&#39;foo&#39;] = &#39;metacls was here&#39;
 return type.__new__(mcs, name, bases, dict)
</pre></div>
</div>
You can of course also override other class methods (or add new methods); for
example defining a custom <a class="reference internal" href="#object.__call__" title="object.__call__"><tt class="xref py py-meth docutils literal">__call__()</tt></a> method in the metaclass allows custom
behavior when the class is called, e.g. not always creating a new instance.
<dl class="data">
<dt id="__metaclass__">
<tt class="descname">__metaclass__</tt><a class="headerlink" href="#__metaclass__" title="Permalink to this definition">¶</a></dt>
<dd>This variable can be any callable accepting arguments for <tt class="docutils literal">name</tt>, <tt class="docutils literal">bases</tt>,
and <tt class="docutils literal">dict</tt>. Upon class creation, the callable is used instead of the built-in
<a class="reference internal" href="../library/functions.html#type" title="type"><tt class="xref py py-func docutils literal">type()</tt></a>.

New in version 2.2.
</dd></dl>

The appropriate metaclass is determined by the following precedence rules:
<ul class="simple">
<li>If <tt class="docutils literal">dict['__metaclass__']</tt> exists, it is used.</li>
<li>Otherwise, if there is at least one base class, its metaclass is used (this
looks for a __class__ attribute first and if not found, uses its type).</li>
<li>Otherwise, if a global variable named __metaclass__ exists, it is used.</li>
<li>Otherwise, the old-style, classic metaclass (types.ClassType) is used.</li>
</ul>
The potential uses for metaclasses are boundless. Some ideas that have been
explored including logging, interface checking, automatic delegation, automatic
property creation, proxies, frameworks, and automatic resource
locking/synchronization.
</div>
<div class="section" id="customizing-instance-and-subclass-checks">
<h3>3.4.4. Customizing instance and subclass checks<a class="headerlink" href="#customizing-instance-and-subclass-checks" title="Permalink to this headline">¶</a></h3>

New in version 2.6.
The following methods are used to override the default behavior of the
<a class="reference internal" href="../library/functions.html#isinstance" title="isinstance"><tt class="xref py py-func docutils literal">isinstance()</tt></a> and <a class="reference internal" href="../library/functions.html#issubclass" title="issubclass"><tt class="xref py py-func docutils literal">issubclass()</tt></a> built-in functions.
In particular, the metaclass <a class="reference internal" href="../library/abc.html#abc.ABCMeta" title="abc.ABCMeta"><tt class="xref py py-class docutils literal">abc.ABCMeta</tt></a> implements these methods in
order to allow the addition of Abstract Base Classes (ABCs) as &#8220;virtual base
classes&#8221; to any class or type (including built-in types), including other
ABCs.
<dl class="method">
<dt id="class.__instancecheck__">
<tt class="descclassname">class.</tt><tt class="descname">__instancecheck__</tt><big>(</big>self, instance<big>)</big><a class="headerlink" href="#class.__instancecheck__" title="Permalink to this definition">¶</a></dt>
<dd>Return true if instance should be considered a (direct or indirect)
instance of class. If defined, called to implement <tt class="docutils literal">isinstance(instance,
class)</tt>.
</dd></dl>

<dl class="method">
<dt id="class.__subclasscheck__">
<tt class="descclassname">class.</tt><tt class="descname">__subclasscheck__</tt><big>(</big>self, subclass<big>)</big><a class="headerlink" href="#class.__subclasscheck__" title="Permalink to this definition">¶</a></dt>
<dd>Return true if subclass should be considered a (direct or indirect)
subclass of class. If defined, called to implement <tt class="docutils literal">issubclass(subclass,
class)</tt>.
</dd></dl>

Note that these methods are looked up on the type (metaclass) of a class. They
cannot be defined as class methods in the actual class. This is consistent with
the lookup of special methods that are called on instances, only in this
case the instance is itself a class.
<div class="admonition-see-also admonition seealso">
See also
<dl class="last docutils">
<dt><a class="pep reference external" href="http://www.python.org/dev/peps/pep-3119">PEP 3119</a> - Introducing Abstract Base Classes</dt>
<dd>Includes the specification for customizing <a class="reference internal" href="../library/functions.html#isinstance" title="isinstance"><tt class="xref py py-func docutils literal">isinstance()</tt></a> and
<a class="reference internal" href="../library/functions.html#issubclass" title="issubclass"><tt class="xref py py-func docutils literal">issubclass()</tt></a> behavior through <tt class="xref py py-meth docutils literal">__instancecheck__()</tt> and
<tt class="xref py py-meth docutils literal">__subclasscheck__()</tt>, with motivation for this functionality in the
context of adding Abstract Base Classes (see the <a class="reference internal" href="../library/abc.html#module-abc" title="abc: Abstract base classes according to PEP 3119."><tt class="xref py py-mod docutils literal">abc</tt></a> module) to the
language.</dd>
</dl>
</div>
</div>
<div class="section" id="emulating-callable-objects">
<h3>3.4.5. Emulating callable objects<a class="headerlink" href="#emulating-callable-objects" title="Permalink to this headline">¶</a></h3>
<dl class="method">
<dt id="object.__call__">
<tt class="descclassname">object.</tt><tt class="descname">__call__</tt><big>(</big>self[, args...]<big>)</big><a class="headerlink" href="#object.__call__" title="Permalink to this definition">¶</a></dt>
<dd>Called when the instance is &#8220;called&#8221; as a function; if this method is defined,
<tt class="docutils literal">x(arg1, arg2, ...)</tt> is a shorthand for <tt class="docutils literal">x.__call__(arg1, arg2, ...)</tt>.
</dd></dl>

</div>
<div class="section" id="emulating-container-types">
<h3>3.4.6. Emulating container types<a class="headerlink" href="#emulating-container-types" title="Permalink to this headline">¶</a></h3>
The following methods can be defined to implement container objects. Containers
usually are sequences (such as lists or tuples) or mappings (like dictionaries),
but can represent other containers as well. The first set of methods is used
either to emulate a sequence or to emulate a mapping; the difference is that for
a sequence, the allowable keys should be the integers k for which <tt class="docutils literal">0 &lt;= k &lt;
N</tt> where N is the length of the sequence, or slice objects, which define a
range of items. (For backwards compatibility, the method <a class="reference internal" href="#object.__getslice__" title="object.__getslice__"><tt class="xref py py-meth docutils literal">__getslice__()</tt></a>
(see below) can also be defined to handle simple, but not extended slices.) It
is also recommended that mappings provide the methods <tt class="xref py py-meth docutils literal">keys()</tt>,
<tt class="xref py py-meth docutils literal">values()</tt>, <tt class="xref py py-meth docutils literal">items()</tt>, <tt class="xref py py-meth docutils literal">has_key()</tt>, <tt class="xref py py-meth docutils literal">get()</tt>, <tt class="xref py py-meth docutils literal">clear()</tt>,
<tt class="xref py py-meth docutils literal">setdefault()</tt>, <tt class="xref py py-meth docutils literal">iterkeys()</tt>, <tt class="xref py py-meth docutils literal">itervalues()</tt>, <tt class="xref py py-meth docutils literal">iteritems()</tt>,
<tt class="xref py py-meth docutils literal">pop()</tt>, <tt class="xref py py-meth docutils literal">popitem()</tt>, <a class="reference internal" href="../library/copy.html#module-copy" title="copy: Shallow and deep copy operations."><tt class="xref py py-meth docutils literal">copy()</tt></a>, and <tt class="xref py py-meth docutils literal">update()</tt> behaving similar
to those for Python&#8217;s standard dictionary objects. The <a class="reference internal" href="../library/userdict.html#module-UserDict" title="UserDict: Class wrapper for dictionary objects."><tt class="xref py py-mod docutils literal">UserDict</tt></a> module
provides a <tt class="xref py py-class docutils literal">DictMixin</tt> class to help create those methods from a base set
of <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a>, <a class="reference internal" href="#object.__setitem__" title="object.__setitem__"><tt class="xref py py-meth docutils literal">__setitem__()</tt></a>, <a class="reference internal" href="#object.__delitem__" title="object.__delitem__"><tt class="xref py py-meth docutils literal">__delitem__()</tt></a>, and
<tt class="xref py py-meth docutils literal">keys()</tt>. Mutable sequences should provide methods <tt class="xref py py-meth docutils literal">append()</tt>,
<tt class="xref py py-meth docutils literal">count()</tt>, <tt class="xref py py-meth docutils literal">index()</tt>, <tt class="xref py py-meth docutils literal">extend()</tt>, <tt class="xref py py-meth docutils literal">insert()</tt>, <tt class="xref py py-meth docutils literal">pop()</tt>,
<tt class="xref py py-meth docutils literal">remove()</tt>, <tt class="xref py py-meth docutils literal">reverse()</tt> and <tt class="xref py py-meth docutils literal">sort()</tt>, like Python standard list
objects. Finally, sequence types should implement addition (meaning
concatenation) and multiplication (meaning repetition) by defining the methods
<a class="reference internal" href="#object.__add__" title="object.__add__"><tt class="xref py py-meth docutils literal">__add__()</tt></a>, <a class="reference internal" href="#object.__radd__" title="object.__radd__"><tt class="xref py py-meth docutils literal">__radd__()</tt></a>, <a class="reference internal" href="#object.__iadd__" title="object.__iadd__"><tt class="xref py py-meth docutils literal">__iadd__()</tt></a>, <a class="reference internal" href="#object.__mul__" title="object.__mul__"><tt class="xref py py-meth docutils literal">__mul__()</tt></a>,
<a class="reference internal" href="#object.__rmul__" title="object.__rmul__"><tt class="xref py py-meth docutils literal">__rmul__()</tt></a> and <a class="reference internal" href="#object.__imul__" title="object.__imul__"><tt class="xref py py-meth docutils literal">__imul__()</tt></a> described below; they should not define
<a class="reference internal" href="#object.__coerce__" title="object.__coerce__"><tt class="xref py py-meth docutils literal">__coerce__()</tt></a> or other numerical operators. It is recommended that both
mappings and sequences implement the <a class="reference internal" href="#object.__contains__" title="object.__contains__"><tt class="xref py py-meth docutils literal">__contains__()</tt></a> method to allow
efficient use of the <tt class="docutils literal">in</tt> operator; for mappings, <tt class="docutils literal">in</tt> should be equivalent
of <tt class="xref py py-meth docutils literal">has_key()</tt>; for sequences, it should search through the values. It is
further recommended that both mappings and sequences implement the
<a class="reference internal" href="#object.__iter__" title="object.__iter__"><tt class="xref py py-meth docutils literal">__iter__()</tt></a> method to allow efficient iteration through the container; for
mappings, <a class="reference internal" href="#object.__iter__" title="object.__iter__"><tt class="xref py py-meth docutils literal">__iter__()</tt></a> should be the same as <tt class="xref py py-meth docutils literal">iterkeys()</tt>; for
sequences, it should iterate through the values.
<dl class="method">
<dt id="object.__len__">
<tt class="descclassname">object.</tt><tt class="descname">__len__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__len__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement the built-in function <a class="reference internal" href="../library/functions.html#len" title="len"><tt class="xref py py-func docutils literal">len()</tt></a>. Should return the length
of the object, an integer <tt class="docutils literal">&gt;=</tt> 0. Also, an object that doesn&#8217;t define a
<a class="reference internal" href="#object.__nonzero__" title="object.__nonzero__"><tt class="xref py py-meth docutils literal">__nonzero__()</tt></a> method and whose <a class="reference internal" href="#object.__len__" title="object.__len__"><tt class="xref py py-meth docutils literal">__len__()</tt></a> method returns zero is
considered to be false in a Boolean context.
</dd></dl>

<dl class="method">
<dt id="object.__getitem__">
<tt class="descclassname">object.</tt><tt class="descname">__getitem__</tt><big>(</big>self, key<big>)</big><a class="headerlink" href="#object.__getitem__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement evaluation of <tt class="docutils literal">self[key]</tt>. For sequence types, the
accepted keys should be integers and slice objects. Note that the special
interpretation of negative indexes (if the class wishes to emulate a sequence
type) is up to the <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a> method. If key is of an inappropriate
type, <a class="reference internal" href="../library/exceptions.html#exceptions.TypeError" title="exceptions.TypeError"><tt class="xref py py-exc docutils literal">TypeError</tt></a> may be raised; if of a value outside the set of indexes
for the sequence (after any special interpretation of negative values),
<a class="reference internal" href="../library/exceptions.html#exceptions.IndexError" title="exceptions.IndexError"><tt class="xref py py-exc docutils literal">IndexError</tt></a> should be raised. For mapping types, if key is missing (not
in the container), <a class="reference internal" href="../library/exceptions.html#exceptions.KeyError" title="exceptions.KeyError"><tt class="xref py py-exc docutils literal">KeyError</tt></a> should be raised.
<div class="admonition note">
Note
<a class="reference internal" href="compound_stmts.html#for"><tt class="xref std std-keyword docutils literal">for</tt></a> loops expect that an <a class="reference internal" href="../library/exceptions.html#exceptions.IndexError" title="exceptions.IndexError"><tt class="xref py py-exc docutils literal">IndexError</tt></a> will be raised for illegal
indexes to allow proper detection of the end of the sequence.
</div>
</dd></dl>

<dl class="method">
<dt id="object.__setitem__">
<tt class="descclassname">object.</tt><tt class="descname">__setitem__</tt><big>(</big>self, key, value<big>)</big><a class="headerlink" href="#object.__setitem__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement assignment to <tt class="docutils literal">self[key]</tt>. Same note as for
<a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a>. This should only be implemented for mappings if the
objects support changes to the values for keys, or if new keys can be added, or
for sequences if elements can be replaced. The same exceptions should be raised
for improper key values as for the <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a> method.
</dd></dl>

<dl class="method">
<dt id="object.__delitem__">
<tt class="descclassname">object.</tt><tt class="descname">__delitem__</tt><big>(</big>self, key<big>)</big><a class="headerlink" href="#object.__delitem__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement deletion of <tt class="docutils literal">self[key]</tt>. Same note as for
<a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a>. This should only be implemented for mappings if the
objects support removal of keys, or for sequences if elements can be removed
from the sequence. The same exceptions should be raised for improper key
values as for the <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a> method.
</dd></dl>

<dl class="method">
<dt id="object.__iter__">
<tt class="descclassname">object.</tt><tt class="descname">__iter__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__iter__" title="Permalink to this definition">¶</a></dt>
<dd>This method is called when an iterator is required for a container. This method
should return a new iterator object that can iterate over all the objects in the
container. For mappings, it should iterate over the keys of the container, and
should also be made available as the method <tt class="xref py py-meth docutils literal">iterkeys()</tt>.
Iterator objects also need to implement this method; they are required to return
themselves. For more information on iterator objects, see <a class="reference internal" href="../library/stdtypes.html#typeiter">Iterator Types</a>.
</dd></dl>

<dl class="method">
<dt id="object.__reversed__">
<tt class="descclassname">object.</tt><tt class="descname">__reversed__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__reversed__" title="Permalink to this definition">¶</a></dt>
<dd>Called (if present) by the <a class="reference internal" href="../library/functions.html#reversed" title="reversed"><tt class="xref py py-func docutils literal">reversed()</tt></a> built-in to implement
reverse iteration. It should return a new iterator object that iterates
over all the objects in the container in reverse order.
If the <a class="reference internal" href="#object.__reversed__" title="object.__reversed__"><tt class="xref py py-meth docutils literal">__reversed__()</tt></a> method is not provided, the <a class="reference internal" href="../library/functions.html#reversed" title="reversed"><tt class="xref py py-func docutils literal">reversed()</tt></a>
built-in will fall back to using the sequence protocol (<a class="reference internal" href="#object.__len__" title="object.__len__"><tt class="xref py py-meth docutils literal">__len__()</tt></a> and
<a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a>). Objects that support the sequence protocol should
only provide <a class="reference internal" href="#object.__reversed__" title="object.__reversed__"><tt class="xref py py-meth docutils literal">__reversed__()</tt></a> if they can provide an implementation
that is more efficient than the one provided by <a class="reference internal" href="../library/functions.html#reversed" title="reversed"><tt class="xref py py-func docutils literal">reversed()</tt></a>.

New in version 2.6.
</dd></dl>

The membership test operators (<a class="reference internal" href="expressions.html#in"><tt class="xref std std-keyword docutils literal">in</tt></a> and <a class="reference internal" href="expressions.html#not-in"><tt class="xref std std-keyword docutils literal">not in</tt></a>) are normally
implemented as an iteration through a sequence. However, container objects can
supply the following special method with a more efficient implementation, which
also does not require the object be a sequence.
<dl class="method">
<dt id="object.__contains__">
<tt class="descclassname">object.</tt><tt class="descname">__contains__</tt><big>(</big>self, item<big>)</big><a class="headerlink" href="#object.__contains__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement membership test operators. Should return true if item
is in self, false otherwise. For mapping objects, this should consider the
keys of the mapping rather than the values or the key-item pairs.
For objects that don&#8217;t define <a class="reference internal" href="#object.__contains__" title="object.__contains__"><tt class="xref py py-meth docutils literal">__contains__()</tt></a>, the membership test first
tries iteration via <a class="reference internal" href="#object.__iter__" title="object.__iter__"><tt class="xref py py-meth docutils literal">__iter__()</tt></a>, then the old sequence iteration
protocol via <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a>, see <a class="reference internal" href="expressions.html#membership-test-details">this section in the language
reference</a>.
</dd></dl>

</div>
<div class="section" id="additional-methods-for-emulation-of-sequence-types">
<h3>3.4.7. Additional methods for emulation of sequence types<a class="headerlink" href="#additional-methods-for-emulation-of-sequence-types" title="Permalink to this headline">¶</a></h3>
The following optional methods can be defined to further emulate sequence
objects. Immutable sequences methods should at most only define
<a class="reference internal" href="#object.__getslice__" title="object.__getslice__"><tt class="xref py py-meth docutils literal">__getslice__()</tt></a>; mutable sequences might define all three methods.
<dl class="method">
<dt id="object.__getslice__">
<tt class="descclassname">object.</tt><tt class="descname">__getslice__</tt><big>(</big>self, i, j<big>)</big><a class="headerlink" href="#object.__getslice__" title="Permalink to this definition">¶</a></dt>
<dd>
Deprecated since version 2.0: Support slice objects as parameters to the <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a> method.
(However, built-in types in CPython currently still implement
<a class="reference internal" href="#object.__getslice__" title="object.__getslice__"><tt class="xref py py-meth docutils literal">__getslice__()</tt></a>. Therefore, you have to override it in derived
classes when implementing slicing.)
Called to implement evaluation of <tt class="docutils literal">self[i:j]</tt>. The returned object should be
of the same type as self. Note that missing i or j in the slice
expression are replaced by zero or <tt class="docutils literal">sys.maxint</tt>, respectively. If negative
indexes are used in the slice, the length of the sequence is added to that
index. If the instance does not implement the <a class="reference internal" href="#object.__len__" title="object.__len__"><tt class="xref py py-meth docutils literal">__len__()</tt></a> method, an
<a class="reference internal" href="../library/exceptions.html#exceptions.AttributeError" title="exceptions.AttributeError"><tt class="xref py py-exc docutils literal">AttributeError</tt></a> is raised. No guarantee is made that indexes adjusted this
way are not still negative. Indexes which are greater than the length of the
sequence are not modified. If no <a class="reference internal" href="#object.__getslice__" title="object.__getslice__"><tt class="xref py py-meth docutils literal">__getslice__()</tt></a> is found, a slice object
is created instead, and passed to <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a> instead.
</dd></dl>

<dl class="method">
<dt id="object.__setslice__">
<tt class="descclassname">object.</tt><tt class="descname">__setslice__</tt><big>(</big>self, i, j, sequence<big>)</big><a class="headerlink" href="#object.__setslice__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement assignment to <tt class="docutils literal">self[i:j]</tt>. Same notes for i and j as
for <a class="reference internal" href="#object.__getslice__" title="object.__getslice__"><tt class="xref py py-meth docutils literal">__getslice__()</tt></a>.
This method is deprecated. If no <a class="reference internal" href="#object.__setslice__" title="object.__setslice__"><tt class="xref py py-meth docutils literal">__setslice__()</tt></a> is found, or for extended
slicing of the form <tt class="docutils literal">self[i:j:k]</tt>, a slice object is created, and passed to
<a class="reference internal" href="#object.__setitem__" title="object.__setitem__"><tt class="xref py py-meth docutils literal">__setitem__()</tt></a>, instead of <a class="reference internal" href="#object.__setslice__" title="object.__setslice__"><tt class="xref py py-meth docutils literal">__setslice__()</tt></a> being called.
</dd></dl>

<dl class="method">
<dt id="object.__delslice__">
<tt class="descclassname">object.</tt><tt class="descname">__delslice__</tt><big>(</big>self, i, j<big>)</big><a class="headerlink" href="#object.__delslice__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement deletion of <tt class="docutils literal">self[i:j]</tt>. Same notes for i and j as for
<a class="reference internal" href="#object.__getslice__" title="object.__getslice__"><tt class="xref py py-meth docutils literal">__getslice__()</tt></a>. This method is deprecated. If no <a class="reference internal" href="#object.__delslice__" title="object.__delslice__"><tt class="xref py py-meth docutils literal">__delslice__()</tt></a> is
found, or for extended slicing of the form <tt class="docutils literal">self[i:j:k]</tt>, a slice object is
created, and passed to <a class="reference internal" href="#object.__delitem__" title="object.__delitem__"><tt class="xref py py-meth docutils literal">__delitem__()</tt></a>, instead of <a class="reference internal" href="#object.__delslice__" title="object.__delslice__"><tt class="xref py py-meth docutils literal">__delslice__()</tt></a>
being called.
</dd></dl>

Notice that these methods are only invoked when a single slice with a single
colon is used, and the slice method is available. For slice operations
involving extended slice notation, or in absence of the slice methods,
<a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a>, <a class="reference internal" href="#object.__setitem__" title="object.__setitem__"><tt class="xref py py-meth docutils literal">__setitem__()</tt></a> or <a class="reference internal" href="#object.__delitem__" title="object.__delitem__"><tt class="xref py py-meth docutils literal">__delitem__()</tt></a> is called with a
slice object as argument.
The following example demonstrate how to make your program or module compatible
with earlier versions of Python (assuming that methods <a class="reference internal" href="#object.__getitem__" title="object.__getitem__"><tt class="xref py py-meth docutils literal">__getitem__()</tt></a>,
<a class="reference internal" href="#object.__setitem__" title="object.__setitem__"><tt class="xref py py-meth docutils literal">__setitem__()</tt></a> and <a class="reference internal" href="#object.__delitem__" title="object.__delitem__"><tt class="xref py py-meth docutils literal">__delitem__()</tt></a> support slice objects as
arguments):
<div class="highlight-python"><div class="highlight"><pre>class MyClass:
 ...
 def __getitem__(self, index):
 ...
 def __setitem__(self, index, value):
 ...
 def __delitem__(self, index):
 ...

if sys.version_info &lt; (2, 0):
 # They won&#39;t be defined if version is at least 2.0 final

def __getslice__(self, i, j):
 return self[max(0, i):max(0, j):]
 def __setslice__(self, i, j, seq):
 self[max(0, i):max(0, j):] = seq
 def __delslice__(self, i, j):
 del self[max(0, i):max(0, j):]
 ...
</pre></div>
</div>
Note the calls to <a class="reference internal" href="../library/functions.html#max" title="max"><tt class="xref py py-func docutils literal">max()</tt></a>; these are necessary because of the handling of
negative indices before the <tt class="xref py py-meth docutils literal">__*slice__()</tt> methods are called. When
negative indexes are used, the <tt class="xref py py-meth docutils literal">__*item__()</tt> methods receive them as
provided, but the <tt class="xref py py-meth docutils literal">__*slice__()</tt> methods get a &#8220;cooked&#8221; form of the index
values. For each negative index value, the length of the sequence is added to
the index before calling the method (which may still result in a negative
index); this is the customary handling of negative indexes by the built-in
sequence types, and the <tt class="xref py py-meth docutils literal">__*item__()</tt> methods are expected to do this as
well. However, since they should already be doing that, negative indexes cannot
be passed in; they must be constrained to the bounds of the sequence before
being passed to the <tt class="xref py py-meth docutils literal">__*item__()</tt> methods. Calling <tt class="docutils literal">max(0, i)</tt>
conveniently returns the proper value.
</div>
<div class="section" id="emulating-numeric-types">
<h3>3.4.8. Emulating numeric types<a class="headerlink" href="#emulating-numeric-types" title="Permalink to this headline">¶</a></h3>
The following methods can be defined to emulate numeric objects. Methods
corresponding to operations that are not supported by the particular kind of
number implemented (e.g., bitwise operations for non-integral numbers) should be
left undefined.
<dl class="method">
<dt id="object.__add__">
<tt class="descclassname">object.</tt><tt class="descname">__add__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__add__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__sub__">
<tt class="descclassname">object.</tt><tt class="descname">__sub__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__sub__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__mul__">
<tt class="descclassname">object.</tt><tt class="descname">__mul__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__mul__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__floordiv__">
<tt class="descclassname">object.</tt><tt class="descname">__floordiv__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__floordiv__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__mod__">
<tt class="descclassname">object.</tt><tt class="descname">__mod__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__mod__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__divmod__">
<tt class="descclassname">object.</tt><tt class="descname">__divmod__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__divmod__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__pow__">
<tt class="descclassname">object.</tt><tt class="descname">__pow__</tt><big>(</big>self, other[, modulo]<big>)</big><a class="headerlink" href="#object.__pow__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__lshift__">
<tt class="descclassname">object.</tt><tt class="descname">__lshift__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__lshift__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rshift__">
<tt class="descclassname">object.</tt><tt class="descname">__rshift__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rshift__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__and__">
<tt class="descclassname">object.</tt><tt class="descname">__and__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__and__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__xor__">
<tt class="descclassname">object.</tt><tt class="descname">__xor__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__xor__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__or__">
<tt class="descclassname">object.</tt><tt class="descname">__or__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__or__" title="Permalink to this definition">¶</a></dt>
<dd>These methods are called to implement the binary arithmetic operations (<tt class="docutils literal">+</tt>,
<tt class="docutils literal">-</tt>, <tt class="docutils literal">*</tt>, <tt class="docutils literal">//</tt>, <tt class="docutils literal">%</tt>, <a class="reference internal" href="../library/functions.html#divmod" title="divmod"><tt class="xref py py-func docutils literal">divmod()</tt></a>, <a class="reference internal" href="../library/functions.html#pow" title="pow"><tt class="xref py py-func docutils literal">pow()</tt></a>, <tt class="docutils literal">**</tt>, <tt class="docutils literal">&lt;&lt;</tt>,
<tt class="docutils literal">&gt;&gt;</tt>, <tt class="docutils literal">&amp;</tt>, <tt class="docutils literal">^</tt>, <tt class="docutils literal">|</tt>). For instance, to evaluate the expression
<tt class="docutils literal">x + y</tt>, where x is an instance of a class that has an <a class="reference internal" href="#object.__add__" title="object.__add__"><tt class="xref py py-meth docutils literal">__add__()</tt></a>
method, <tt class="docutils literal">x.__add__(y)</tt> is called. The <a class="reference internal" href="#object.__divmod__" title="object.__divmod__"><tt class="xref py py-meth docutils literal">__divmod__()</tt></a> method should be the
equivalent to using <a class="reference internal" href="#object.__floordiv__" title="object.__floordiv__"><tt class="xref py py-meth docutils literal">__floordiv__()</tt></a> and <a class="reference internal" href="#object.__mod__" title="object.__mod__"><tt class="xref py py-meth docutils literal">__mod__()</tt></a>; it should not be
related to <a class="reference internal" href="#object.__truediv__" title="object.__truediv__"><tt class="xref py py-meth docutils literal">__truediv__()</tt></a> (described below). Note that <a class="reference internal" href="#object.__pow__" title="object.__pow__"><tt class="xref py py-meth docutils literal">__pow__()</tt></a>
should be defined to accept an optional third argument if the ternary version of
the built-in <a class="reference internal" href="../library/functions.html#pow" title="pow"><tt class="xref py py-func docutils literal">pow()</tt></a> function is to be supported.
If one of those methods does not support the operation with the supplied
arguments, it should return <tt class="docutils literal">NotImplemented</tt>.
</dd></dl>

<dl class="method">
<dt id="object.__div__">
<tt class="descclassname">object.</tt><tt class="descname">__div__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__div__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__truediv__">
<tt class="descclassname">object.</tt><tt class="descname">__truediv__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__truediv__" title="Permalink to this definition">¶</a></dt>
<dd>The division operator (<tt class="docutils literal">/</tt>) is implemented by these methods. The
<a class="reference internal" href="#object.__truediv__" title="object.__truediv__"><tt class="xref py py-meth docutils literal">__truediv__()</tt></a> method is used when <tt class="docutils literal">__future__.division</tt> is in effect,
otherwise <a class="reference internal" href="#object.__div__" title="object.__div__"><tt class="xref py py-meth docutils literal">__div__()</tt></a> is used. If only one of these two methods is defined,
the object will not support division in the alternate context; <a class="reference internal" href="../library/exceptions.html#exceptions.TypeError" title="exceptions.TypeError"><tt class="xref py py-exc docutils literal">TypeError</tt></a>
will be raised instead.
</dd></dl>

<dl class="method">
<dt id="object.__radd__">
<tt class="descclassname">object.</tt><tt class="descname">__radd__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__radd__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rsub__">
<tt class="descclassname">object.</tt><tt class="descname">__rsub__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rsub__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rmul__">
<tt class="descclassname">object.</tt><tt class="descname">__rmul__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rmul__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rdiv__">
<tt class="descclassname">object.</tt><tt class="descname">__rdiv__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rdiv__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rtruediv__">
<tt class="descclassname">object.</tt><tt class="descname">__rtruediv__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rtruediv__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rfloordiv__">
<tt class="descclassname">object.</tt><tt class="descname">__rfloordiv__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rfloordiv__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rmod__">
<tt class="descclassname">object.</tt><tt class="descname">__rmod__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rmod__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rdivmod__">
<tt class="descclassname">object.</tt><tt class="descname">__rdivmod__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rdivmod__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rpow__">
<tt class="descclassname">object.</tt><tt class="descname">__rpow__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rpow__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rlshift__">
<tt class="descclassname">object.</tt><tt class="descname">__rlshift__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rlshift__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rrshift__">
<tt class="descclassname">object.</tt><tt class="descname">__rrshift__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rrshift__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rand__">
<tt class="descclassname">object.</tt><tt class="descname">__rand__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rand__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__rxor__">
<tt class="descclassname">object.</tt><tt class="descname">__rxor__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__rxor__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__ror__">
<tt class="descclassname">object.</tt><tt class="descname">__ror__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__ror__" title="Permalink to this definition">¶</a></dt>
<dd>These methods are called to implement the binary arithmetic operations (<tt class="docutils literal">+</tt>,
<tt class="docutils literal">-</tt>, <tt class="docutils literal">*</tt>, <tt class="docutils literal">/</tt>, <tt class="docutils literal">%</tt>, <a class="reference internal" href="../library/functions.html#divmod" title="divmod"><tt class="xref py py-func docutils literal">divmod()</tt></a>, <a class="reference internal" href="../library/functions.html#pow" title="pow"><tt class="xref py py-func docutils literal">pow()</tt></a>, <tt class="docutils literal">**</tt>, <tt class="docutils literal">&lt;&lt;</tt>, <tt class="docutils literal">&gt;&gt;</tt>,
<tt class="docutils literal">&amp;</tt>, <tt class="docutils literal">^</tt>, <tt class="docutils literal">|</tt>) with reflected (swapped) operands. These functions are
only called if the left operand does not support the corresponding operation and
the operands are of different types. <a class="footnote-reference" href="#id6" id="id3">[2]</a> For instance, to evaluate the
expression <tt class="docutils literal">x - y</tt>, where y is an instance of a class that has an
<a class="reference internal" href="#object.__rsub__" title="object.__rsub__"><tt class="xref py py-meth docutils literal">__rsub__()</tt></a> method, <tt class="docutils literal">y.__rsub__(x)</tt> is called if <tt class="docutils literal">x.__sub__(y)</tt> returns
NotImplemented.
Note that ternary <a class="reference internal" href="../library/functions.html#pow" title="pow"><tt class="xref py py-func docutils literal">pow()</tt></a> will not try calling <a class="reference internal" href="#object.__rpow__" title="object.__rpow__"><tt class="xref py py-meth docutils literal">__rpow__()</tt></a> (the
coercion rules would become too complicated).
<div class="admonition note">
Note
If the right operand&#8217;s type is a subclass of the left operand&#8217;s type and that
subclass provides the reflected method for the operation, this method will be
called before the left operand&#8217;s non-reflected method. This behavior allows
subclasses to override their ancestors&#8217; operations.
</div>
</dd></dl>

<dl class="method">
<dt id="object.__iadd__">
<tt class="descclassname">object.</tt><tt class="descname">__iadd__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__iadd__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__isub__">
<tt class="descclassname">object.</tt><tt class="descname">__isub__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__isub__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__imul__">
<tt class="descclassname">object.</tt><tt class="descname">__imul__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__imul__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__idiv__">
<tt class="descclassname">object.</tt><tt class="descname">__idiv__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__idiv__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__itruediv__">
<tt class="descclassname">object.</tt><tt class="descname">__itruediv__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__itruediv__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__ifloordiv__">
<tt class="descclassname">object.</tt><tt class="descname">__ifloordiv__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__ifloordiv__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__imod__">
<tt class="descclassname">object.</tt><tt class="descname">__imod__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__imod__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__ipow__">
<tt class="descclassname">object.</tt><tt class="descname">__ipow__</tt><big>(</big>self, other[, modulo]<big>)</big><a class="headerlink" href="#object.__ipow__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__ilshift__">
<tt class="descclassname">object.</tt><tt class="descname">__ilshift__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__ilshift__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__irshift__">
<tt class="descclassname">object.</tt><tt class="descname">__irshift__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__irshift__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__iand__">
<tt class="descclassname">object.</tt><tt class="descname">__iand__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__iand__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__ixor__">
<tt class="descclassname">object.</tt><tt class="descname">__ixor__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__ixor__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__ior__">
<tt class="descclassname">object.</tt><tt class="descname">__ior__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__ior__" title="Permalink to this definition">¶</a></dt>
<dd>These methods are called to implement the augmented arithmetic assignments
(<tt class="docutils literal">+=</tt>, <tt class="docutils literal">-=</tt>, <tt class="docutils literal">*=</tt>, <tt class="docutils literal">/=</tt>, <tt class="docutils literal">//=</tt>, <tt class="docutils literal">%=</tt>, <tt class="docutils literal">**=</tt>, <tt class="docutils literal">&lt;&lt;=</tt>, <tt class="docutils literal">&gt;&gt;=</tt>,
<tt class="docutils literal">&amp;=</tt>, <tt class="docutils literal">^=</tt>, <tt class="docutils literal">|=</tt>). These methods should attempt to do the operation
in-place (modifying self) and return the result (which could be, but does
not have to be, self). If a specific method is not defined, the augmented
assignment falls back to the normal methods. For instance, to execute the
statement <tt class="docutils literal">x += y</tt>, where x is an instance of a class that has an
<a class="reference internal" href="#object.__iadd__" title="object.__iadd__"><tt class="xref py py-meth docutils literal">__iadd__()</tt></a> method, <tt class="docutils literal">x.__iadd__(y)</tt> is called. If x is an instance
of a class that does not define a <a class="reference internal" href="#object.__iadd__" title="object.__iadd__"><tt class="xref py py-meth docutils literal">__iadd__()</tt></a> method, <tt class="docutils literal">x.__add__(y)</tt>
and <tt class="docutils literal">y.__radd__(x)</tt> are considered, as with the evaluation of <tt class="docutils literal">x + y</tt>.
</dd></dl>

<dl class="method">
<dt id="object.__neg__">
<tt class="descclassname">object.</tt><tt class="descname">__neg__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__neg__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__pos__">
<tt class="descclassname">object.</tt><tt class="descname">__pos__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__pos__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__abs__">
<tt class="descclassname">object.</tt><tt class="descname">__abs__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__abs__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__invert__">
<tt class="descclassname">object.</tt><tt class="descname">__invert__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__invert__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement the unary arithmetic operations (<tt class="docutils literal">-</tt>, <tt class="docutils literal">+</tt>, <a class="reference internal" href="../library/functions.html#abs" title="abs"><tt class="xref py py-func docutils literal">abs()</tt></a>
and <tt class="docutils literal">~</tt>).
</dd></dl>

<dl class="method">
<dt id="object.__complex__">
<tt class="descclassname">object.</tt><tt class="descname">__complex__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__complex__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__int__">
<tt class="descclassname">object.</tt><tt class="descname">__int__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__int__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__long__">
<tt class="descclassname">object.</tt><tt class="descname">__long__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__long__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__float__">
<tt class="descclassname">object.</tt><tt class="descname">__float__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__float__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement the built-in functions <a class="reference internal" href="../library/functions.html#complex" title="complex"><tt class="xref py py-func docutils literal">complex()</tt></a>, <a class="reference internal" href="../library/functions.html#int" title="int"><tt class="xref py py-func docutils literal">int()</tt></a>,
<a class="reference internal" href="../library/functions.html#long" title="long"><tt class="xref py py-func docutils literal">long()</tt></a>, and <a class="reference internal" href="../library/functions.html#float" title="float"><tt class="xref py py-func docutils literal">float()</tt></a>. Should return a value of the appropriate type.
</dd></dl>

<dl class="method">
<dt id="object.__oct__">
<tt class="descclassname">object.</tt><tt class="descname">__oct__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__oct__" title="Permalink to this definition">¶</a></dt>
<dt id="object.__hex__">
<tt class="descclassname">object.</tt><tt class="descname">__hex__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__hex__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement the built-in functions <a class="reference internal" href="../library/functions.html#oct" title="oct"><tt class="xref py py-func docutils literal">oct()</tt></a> and <a class="reference internal" href="../library/functions.html#hex" title="hex"><tt class="xref py py-func docutils literal">hex()</tt></a>. Should
return a string value.
</dd></dl>

<dl class="method">
<dt id="object.__index__">
<tt class="descclassname">object.</tt><tt class="descname">__index__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__index__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement <a class="reference internal" href="../library/operator.html#operator.index" title="operator.index"><tt class="xref py py-func docutils literal">operator.index()</tt></a>. Also called whenever Python needs
an integer object (such as in slicing). Must return an integer (int or long).

New in version 2.5.
</dd></dl>

<dl class="method">
<dt id="object.__coerce__">
<tt class="descclassname">object.</tt><tt class="descname">__coerce__</tt><big>(</big>self, other<big>)</big><a class="headerlink" href="#object.__coerce__" title="Permalink to this definition">¶</a></dt>
<dd>Called to implement &#8220;mixed-mode&#8221; numeric arithmetic. Should either return a
2-tuple containing self and other converted to a common numeric type, or
<tt class="docutils literal">None</tt> if conversion is impossible. When the common type would be the type of
<tt class="docutils literal">other</tt>, it is sufficient to return <tt class="docutils literal">None</tt>, since the interpreter will also
ask the other object to attempt a coercion (but sometimes, if the implementation
of the other type cannot be changed, it is useful to do the conversion to the
other type here). A return value of <tt class="docutils literal">NotImplemented</tt> is equivalent to
returning <tt class="docutils literal">None</tt>.
</dd></dl>

</div>
<div class="section" id="coercion-rules">
<h3>3.4.9. Coercion rules<a class="headerlink" href="#coercion-rules" title="Permalink to this headline">¶</a></h3>
This section used to document the rules for coercion. As the language has
evolved, the coercion rules have become hard to document precisely; documenting
what one version of one particular implementation does is undesirable. Instead,
here are some informal guidelines regarding coercion. In Python 3, coercion
will not be supported.
<ul>
<li>If the left operand of a % operator is a string or Unicode object, no coercion
takes place and the string formatting operation is invoked instead.
</li>
<li>It is no longer recommended to define a coercion operation. Mixed-mode
operations on types that don&#8217;t define coercion pass the original arguments to
the operation.
</li>
<li>New-style classes (those derived from <a class="reference internal" href="../library/functions.html#object" title="object"><tt class="xref py py-class docutils literal">object</tt></a>) never invoke the
<a class="reference internal" href="#object.__coerce__" title="object.__coerce__"><tt class="xref py py-meth docutils literal">__coerce__()</tt></a> method in response to a binary operator; the only time
<a class="reference internal" href="#object.__coerce__" title="object.__coerce__"><tt class="xref py py-meth docutils literal">__coerce__()</tt></a> is invoked is when the built-in function <a class="reference internal" href="../library/functions.html#coerce" title="coerce"><tt class="xref py py-func docutils literal">coerce()</tt></a> is
called.
</li>
<li>For most intents and purposes, an operator that returns <tt class="docutils literal">NotImplemented</tt> is
treated the same as one that is not implemented at all.
</li>
<li>Below, <tt class="xref py py-meth docutils literal">__op__()</tt> and <tt class="xref py py-meth docutils literal">__rop__()</tt> are used to signify the generic method
names corresponding to an operator; <tt class="xref py py-meth docutils literal">__iop__()</tt> is used for the
corresponding in-place operator. For example, for the operator &#8216;<tt class="docutils literal">+</tt>&#8216;,
<a class="reference internal" href="#object.__add__" title="object.__add__"><tt class="xref py py-meth docutils literal">__add__()</tt></a> and <a class="reference internal" href="#object.__radd__" title="object.__radd__"><tt class="xref py py-meth docutils literal">__radd__()</tt></a> are used for the left and right variant of
the binary operator, and <a class="reference internal" href="#object.__iadd__" title="object.__iadd__"><tt class="xref py py-meth docutils literal">__iadd__()</tt></a> for the in-place variant.
</li>
<li>For objects x and y, first <tt class="docutils literal">x.__op__(y)</tt> is tried. If this is not
implemented or returns <tt class="docutils literal">NotImplemented</tt>, <tt class="docutils literal">y.__rop__(x)</tt> is tried. If this
is also not implemented or returns <tt class="docutils literal">NotImplemented</tt>, a <a class="reference internal" href="../library/exceptions.html#exceptions.TypeError" title="exceptions.TypeError"><tt class="xref py py-exc docutils literal">TypeError</tt></a>
exception is raised. But see the following exception:
</li>
<li>Exception to the previous item: if the left operand is an instance of a built-in
type or a new-style class, and the right operand is an instance of a proper
subclass of that type or class and overrides the base&#8217;s <tt class="xref py py-meth docutils literal">__rop__()</tt> method,
the right operand&#8217;s <tt class="xref py py-meth docutils literal">__rop__()</tt> method is tried before the left operand&#8217;s
<tt class="xref py py-meth docutils literal">__op__()</tt> method.
This is done so that a subclass can completely override binary operators.
Otherwise, the left operand&#8217;s <tt class="xref py py-meth docutils literal">__op__()</tt> method would always accept the
right operand: when an instance of a given class is expected, an instance of a
subclass of that class is always acceptable.
</li>
<li>When either operand type defines a coercion, this coercion is called before that
type&#8217;s <tt class="xref py py-meth docutils literal">__op__()</tt> or <tt class="xref py py-meth docutils literal">__rop__()</tt> method is called, but no sooner. If
the coercion returns an object of a different type for the operand whose
coercion is invoked, part of the process is redone using the new object.
</li>
<li>When an in-place operator (like &#8216;<tt class="docutils literal">+=</tt>&#8216;) is used, if the left operand
implements <tt class="xref py py-meth docutils literal">__iop__()</tt>, it is invoked without any coercion. When the
operation falls back to <tt class="xref py py-meth docutils literal">__op__()</tt> and/or <tt class="xref py py-meth docutils literal">__rop__()</tt>, the normal
coercion rules apply.
</li>
<li>In <tt class="docutils literal">x + y</tt>, if x is a sequence that implements sequence concatenation,
sequence concatenation is invoked.
</li>
<li>In <tt class="docutils literal">x * y</tt>, if one operand is a sequence that implements sequence
repetition, and the other is an integer (<a class="reference internal" href="../library/functions.html#int" title="int"><tt class="xref py py-class docutils literal">int</tt></a> or <a class="reference internal" href="../library/functions.html#long" title="long"><tt class="xref py py-class docutils literal">long</tt></a>),
sequence repetition is invoked.
</li>
<li>Rich comparisons (implemented by methods <a class="reference internal" href="#object.__eq__" title="object.__eq__"><tt class="xref py py-meth docutils literal">__eq__()</tt></a> and so on) never use
coercion. Three-way comparison (implemented by <a class="reference internal" href="#object.__cmp__" title="object.__cmp__"><tt class="xref py py-meth docutils literal">__cmp__()</tt></a>) does use
coercion under the same conditions as other binary operations use it.
</li>
<li>In the current implementation, the built-in numeric types <a class="reference internal" href="../library/functions.html#int" title="int"><tt class="xref py py-class docutils literal">int</tt></a>,
<a class="reference internal" href="../library/functions.html#long" title="long"><tt class="xref py py-class docutils literal">long</tt></a>, <a class="reference internal" href="../library/functions.html#float" title="float"><tt class="xref py py-class docutils literal">float</tt></a>, and <a class="reference internal" href="../library/functions.html#complex" title="complex"><tt class="xref py py-class docutils literal">complex</tt></a> do not use coercion.
All these types implement a <a class="reference internal" href="#object.__coerce__" title="object.__coerce__"><tt class="xref py py-meth docutils literal">__coerce__()</tt></a> method, for use by the built-in
<a class="reference internal" href="../library/functions.html#coerce" title="coerce"><tt class="xref py py-func docutils literal">coerce()</tt></a> function.

Changed in version 2.7.
</li>
</ul>
</div>
<div class="section" id="with-statement-context-managers">
<h3>3.4.10. With Statement Context Managers<a class="headerlink" href="#with-statement-context-managers" title="Permalink to this headline">¶</a></h3>

New in version 2.5.
A context manager is an object that defines the runtime context to be
established when executing a <a class="reference internal" href="compound_stmts.html#with"><tt class="xref std std-keyword docutils literal">with</tt></a> statement. The context manager
handles the entry into, and the exit from, the desired runtime context for the
execution of the block of code. Context managers are normally invoked using the
<a class="reference internal" href="compound_stmts.html#with"><tt class="xref std std-keyword docutils literal">with</tt></a> statement (described in section <a class="reference internal" href="compound_stmts.html#with">The with statement</a>), but can also be
used by directly invoking their methods.
Typical uses of context managers include saving and restoring various kinds of
global state, locking and unlocking resources, closing opened files, etc.
For more information on context managers, see <a class="reference internal" href="../library/stdtypes.html#typecontextmanager">Context Manager Types</a>.
<dl class="method">
<dt id="object.__enter__">
<tt class="descclassname">object.</tt><tt class="descname">__enter__</tt><big>(</big>self<big>)</big><a class="headerlink" href="#object.__enter__" title="Permalink to this definition">¶</a></dt>
<dd>Enter the runtime context related to this object. The <a class="reference internal" href="compound_stmts.html#with"><tt class="xref std std-keyword docutils literal">with</tt></a> statement
will bind this method&#8217;s return value to the target(s) specified in the
<a class="reference internal" href="compound_stmts.html#as"><tt class="xref std std-keyword docutils literal">as</tt></a> clause of the statement, if any.
</dd></dl>

<dl class="method">
<dt id="object.__exit__">
<tt class="descclassname">object.</tt><tt class="descname">__exit__</tt><big>(</big>self, exc_type, exc_value, traceback<big>)</big><a class="headerlink" href="#object.__exit__" title="Permalink to this definition">¶</a></dt>
<dd>Exit the runtime context related to this object. The parameters describe the
exception that caused the context to be exited. If the context was exited
without an exception, all three arguments will be <a class="reference internal" href="../library/constants.html#None" title="None"><tt class="xref py py-const docutils literal">None</tt></a>.
If an exception is supplied, and the method wishes to suppress the exception
(i.e., prevent it from being propagated), it should return a true value.
Otherwise, the exception will be processed normally upon exit from this method.
Note that <a class="reference internal" href="#object.__exit__" title="object.__exit__"><tt class="xref py py-meth docutils literal">__exit__()</tt></a> methods should not reraise the passed-in exception;
this is the caller&#8217;s responsibility.
</dd></dl>

<div class="admonition-see-also admonition seealso">
See also
<dl class="last docutils">
<dt><a class="pep reference external" href="http://www.python.org/dev/peps/pep-0343">PEP 0343</a> - The &#8220;with&#8221; statement</dt>
<dd>The specification, background, and examples for the Python <a class="reference internal" href="compound_stmts.html#with"><tt class="xref std std-keyword docutils literal">with</tt></a>
statement.</dd>
</dl>
</div>
</div>
<div class="section" id="special-method-lookup-for-old-style-classes">
<h3>3.4.11. Special method lookup for old-style classes<a class="headerlink" href="#special-method-lookup-for-old-style-classes" title="Permalink to this headline">¶</a></h3>
For old-style classes, special methods are always looked up in exactly the
same way as any other method or attribute. This is the case regardless of
whether the method is being looked up explicitly as in <tt class="docutils literal">x.__getitem__(i)</tt>
or implicitly as in <tt class="docutils literal">x[i]</tt>.
This behaviour means that special methods may exhibit different behaviour
for different instances of a single old-style class if the appropriate
special attributes are set differently:
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; class C:
... pass
...
&gt;&gt;&gt; c1 = C()
&gt;&gt;&gt; c2 = C()
&gt;&gt;&gt; c1.__len__ = lambda: 5
&gt;&gt;&gt; c2.__len__ = lambda: 9
&gt;&gt;&gt; len(c1)
5
&gt;&gt;&gt; len(c2)
9
</pre></div>
</div>
</div>
<div class="section" id="special-method-lookup-for-new-style-classes">
<h3>3.4.12. Special method lookup for new-style classes<a class="headerlink" href="#special-method-lookup-for-new-style-classes" title="Permalink to this headline">¶</a></h3>
For new-style classes, implicit invocations of special methods are only guaranteed
to work correctly if defined on an object&#8217;s type, not in the object&#8217;s instance
dictionary. That behaviour is the reason why the following code raises an
exception (unlike the equivalent example with old-style classes):
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; class C(object):
... pass
...
&gt;&gt;&gt; c = C()
&gt;&gt;&gt; c.__len__ = lambda: 5
&gt;&gt;&gt; len(c)
Traceback (most recent call last):
 File &quot;&lt;stdin&gt;&quot;, line 1, in &lt;module&gt;
TypeError: object of type &#39;C&#39; has no len()
</pre></div>
</div>
The rationale behind this behaviour lies with a number of special methods such
as <a class="reference internal" href="#object.__hash__" title="object.__hash__"><tt class="xref py py-meth docutils literal">__hash__()</tt></a> and <a class="reference internal" href="#object.__repr__" title="object.__repr__"><tt class="xref py py-meth docutils literal">__repr__()</tt></a> that are implemented by all objects,
including type objects. If the implicit lookup of these methods used the
conventional lookup process, they would fail when invoked on the type object
itself:
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; 1 .__hash__() == hash(1)
True
&gt;&gt;&gt; int.__hash__() == hash(int)
Traceback (most recent call last):
 File &quot;&lt;stdin&gt;&quot;, line 1, in &lt;module&gt;
TypeError: descriptor &#39;__hash__&#39; of &#39;int&#39; object needs an argument
</pre></div>
</div>
Incorrectly attempting to invoke an unbound method of a class in this way is
sometimes referred to as &#8216;metaclass confusion&#8217;, and is avoided by bypassing
the instance when looking up special methods:
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; type(1).__hash__(1) == hash(1)
True
&gt;&gt;&gt; type(int).__hash__(int) == hash(int)
True
</pre></div>
</div>
In addition to bypassing any instance attributes in the interest of
correctness, implicit special method lookup generally also bypasses the
<a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><tt class="xref py py-meth docutils literal">__getattribute__()</tt></a> method even of the object&#8217;s metaclass:
<div class="highlight-python"><div class="highlight"><pre>&gt;&gt;&gt; class Meta(type):
... def __getattribute__(*args):
... print &quot;Metaclass getattribute invoked&quot;
... return type.__getattribute__(*args)
...
&gt;&gt;&gt; class C(object):
... __metaclass__ = Meta
... def __len__(self):
... return 10
... def __getattribute__(*args):
... print &quot;Class getattribute invoked&quot;
... return object.__getattribute__(*args)
...
&gt;&gt;&gt; c = C()
&gt;&gt;&gt; c.__len__() # Explicit lookup via instance
Class getattribute invoked
10
&gt;&gt;&gt; type(c).__len__(c) # Explicit lookup via type
Metaclass getattribute invoked
10
&gt;&gt;&gt; len(c) # Implicit lookup
10
</pre></div>
</div>
Bypassing the <a class="reference internal" href="#object.__getattribute__" title="object.__getattribute__"><tt class="xref py py-meth docutils literal">__getattribute__()</tt></a> machinery in this fashion
provides significant scope for speed optimisations within the
interpreter, at the cost of some flexibility in the handling of
special methods (the special method must be set on the class
object itself in order to be consistently invoked by the interpreter).
Footnotes
<table class="docutils footnote" frame="void" id="id5" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id1">[1]</a></td><td>It is possible in some cases to change an object&#8217;s type, under certain
controlled conditions. It generally isn&#8217;t a good idea though, since it can
lead to some very strange behaviour if it is handled incorrectly.</td></tr>
</tbody>
</table>
<table class="docutils footnote" frame="void" id="id6" rules="none">
<colgroup><col class="label" /><col /></colgroup>
<tbody valign="top">
<tr><td class="label"><a class="fn-backref" href="#id3">[2]</a></td><td>For operands of the same type, it is assumed that if the non-reflected method
(such as <a class="reference internal" href="#object.__add__" title="object.__add__"><tt class="xref py py-meth docutils literal">__add__()</tt></a>) fails the operation is not supported, which is why the
reflected method is not called.</td></tr>
</tbody>
</table>
</div>
</div>
</div>

</div>
 </div>
 </div>
 <div class="sphinxsidebar">
 <div class="sphinxsidebarwrapper">
 <h3><a href="../contents.html">Table Of Contents</a></h3>
 <ul>
<li><a class="reference internal" href="#">3. Data model</a><ul>
<li><a class="reference internal" href="#objects-values-and-types">3.1. Objects, values and types</a></li>
<li><a class="reference internal" href="#the-standard-type-hierarchy">3.2. The standard type hierarchy</a></li>
<li><a class="reference internal" href="#new-style-and-classic-classes">3.3. New-style and classic classes</a></li>
<li><a class="reference internal" href="#special-method-names">3.4. Special method names</a><ul>
<li><a class="reference internal" href="#basic-customization">3.4.1. Basic customization</a></li>
<li><a class="reference internal" href="#customizing-attribute-access">3.4.2. Customizing attribute access</a><ul>
<li><a class="reference internal" href="#more-attribute-access-for-new-style-classes">3.4.2.1. More attribute access for new-style classes</a></li>
<li><a class="reference internal" href="#implementing-descriptors">3.4.2.2. Implementing Descriptors</a></li>
<li><a class="reference internal" href="#invoking-descriptors">3.4.2.3. Invoking Descriptors</a></li>
<li><a class="reference internal" href="#slots">3.4.2.4. __slots__</a></li>
</ul>
</li>
<li><a class="reference internal" href="#customizing-class-creation">3.4.3. Customizing class creation</a></li>
<li><a class="reference internal" href="#customizing-instance-and-subclass-checks">3.4.4. Customizing instance and subclass checks</a></li>
<li><a class="reference internal" href="#emulating-callable-objects">3.4.5. Emulating callable objects</a></li>
<li><a class="reference internal" href="#emulating-container-types">3.4.6. Emulating container types</a></li>
<li><a class="reference internal" href="#additional-methods-for-emulation-of-sequence-types">3.4.7. Additional methods for emulation of sequence types</a></li>
<li><a class="reference internal" href="#emulating-numeric-types">3.4.8. Emulating numeric types</a></li>
<li><a class="reference internal" href="#coercion-rules">3.4.9. Coercion rules</a></li>
<li><a class="reference internal" href="#with-statement-context-managers">3.4.10. With Statement Context Managers</a></li>
<li><a class="reference internal" href="#special-method-lookup-for-old-style-classes">3.4.11. Special method lookup for old-style classes</a></li>
<li><a class="reference internal" href="#special-method-lookup-for-new-style-classes">3.4.12. Special method lookup for new-style classes</a></li>
</ul>
</li>
</ul>
</li>
</ul>

<h4>Previous topic</h4>
 <a href="lexical_analysis.html"
 title="previous chapter">2. Lexical analysis</a>
 <h4>Next topic</h4>
 <a href="executionmodel.html"
 title="next chapter">4. Execution model</a>
<h3>This Page</h3>
<ul class="this-page-menu">
 <li><a href="../bugs.html">Report a Bug</a></li>
 <li><a href="../_sources/reference/datamodel.txt"
 rel="nofollow">Show Source</a></li>
</ul>

<div id="searchbox" style="display: none">
 <h3>Quick search</h3>
 <form class="search" action="../search.html" method="get">
 <input type="text" name="q" />
 <input type="submit" value="Go" />
 <input type="hidden" name="check_keywords" value="yes" />
 <input type="hidden" name="area" value="default" />
 </form>
 
 Enter search terms or a module, class or function name.
 
</div>
<script type="text/javascript">$('#searchbox').show(0);</script>
 </div>
 </div>
 <div class="clearer"></div>
 </div>
 <div class="related">
 <h3>Navigation</h3>
 <ul>
 <li class="right" style="margin-right: 10px">
 <a href="../genindex.html" title="General Index"
 >index</a></li>
 <li class="right" >
 <a href="../py-modindex.html" title="Python Module Index"
 >modules</a> |</li>
 <li class="right" >
 <a href="executionmodel.html" title="4. Execution model"
 >next</a> |</li>
 <li class="right" >
 <a href="lexical_analysis.html" title="2. Lexical analysis"
 >previous</a> |</li>
 <li><img src="../_static/py.png" alt=""
 style="vertical-align: middle; margin-top: -1px"/></li>
 <li><a href="http://www.python.org/">Python</a> &raquo;</li>
 <li>
 <a href="../index.html">Python 2.7.5 documentation</a> &raquo;
 </li>

<li><a href="index.html" >The Python Language Reference</a> &raquo;</li> 
 </ul>
 </div>
 <div class="footer">
 &copy; <a href="../copyright.html">Copyright</a> 1990-2020, Python Software Foundation.
 
 The Python Software Foundation is a non-profit corporation.
 <a href="http://www.python.org/psf/donations/">Please donate.</a>
 
 Last updated on Oct 13, 2020.
 <a href="../bugs.html">Found a bug</a>?
 
 Created using <a href="http://sphinx.pocoo.org/">Sphinx</a> 1.1.3.
 </div>

</body>
</html>