3. Data model

3.1. Objects, values and types

Objects are Python’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’s model of a “stored program computer,” code is also represented by objects.)

Every object has an identity, a type and a value. An object’s identity never changes once it has been created; you may think of it as the object’s address in memory. The ‘is‘ operator compares the identity of two objects; the id() function returns an integer representing its identity (currently implemented as its address). An object’s type is also unchangeable. [1] An object’s type determines the operations that the object supports (e.g., “does it have a length?”) and also defines the possible values for objects of that type. The type() function returns an object’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’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’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 — it is a matter of implementation quality how garbage collection is implemented, as long as no objects are collected that are still reachable.

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 gc module for information on controlling the collection of cyclic garbage. Other implementations act differently and CPython may change.

Note that the use of the implementation’s tracing or debugging facilities may keep objects alive that would normally be collectable. Also note that catching an exception with a ‘try...except‘ statement may keep objects alive.

Some objects contain references to “external” 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 close() method. Programs are strongly recommended to explicitly close such objects. The ‘try...finally‘ 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’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 a = 1; b = 1, a and b may or may not refer to the same object with the value one, depending on the implementation, but after c = []; d = [], c and d are guaranteed to refer to two different, unique, newly created empty lists. (Note that c = d = [] assigns the same object to both c and d.)

3.2. The standard type hierarchy

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 ‘special attributes.’ These are attributes that provide access to the implementation and are not intended for general use. Their definition may change in the future.


This type has a single value. There is a single object with this value. This object is accessed through the built-in name None. It is used to signify the absence of a value in many situations, e.g., it is returned from functions that don’t explicitly return anything. Its truth value is false.


This type has a single value. There is a single object with this value. This object is accessed through the built-in name NotImplemented. 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.


This type has a single value. There is a single object with this value. This object is accessed through the built-in name Ellipsis. It is used to indicate the presence of the ... syntax in a slice. Its truth value is true.


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:


These represent elements from the mathematical set of integers (positive and negative).

There are three types of integers:

Plain integers

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 OverflowError is raised instead). For the purpose of shift and mask operations, integers are assumed to have a binary, 2’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).

Long integers

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’s complement which gives the illusion of an infinite string of sign bits extending to the left.


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 "False" or "True" are returned, respectively.

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.

numbers.Real (float)

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.


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 z can be retrieved through the read-only attributes z.real and z.imag.


These represent finite ordered sets indexed by non-negative numbers. The built-in function len() 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 a[i].

Sequences also support slicing: a[i:j] selects all items with index k such that i <= k < 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 “extended slicing” with a third “step” parameter: a[i:j:k] selects all items of a with index x where x = i + n*k, n >= 0 and i <= x < j.

Sequences are distinguished according to their mutability:

Immutable sequences

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:


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 chr() and ord() 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 chr() and ord() implement a mapping between ASCII and EBCDIC, and string comparison preserves the ASCII order. Or perhaps someone can propose a better rule?)


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 sys.maxunicode, 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 unichr() and ord() 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 encode() and the built-in function unicode().


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 ‘singleton’) 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.

Mutable sequences

Mutable sequences can be changed after they are created. The subscription and slicing notations can be used as the target of assignment and del (delete) statements.

There are currently two intrinsic mutable sequence types:


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.)

Byte Arrays

A bytearray object is a mutable array. They are created by the built-in bytearray() constructor. Aside from being mutable (and hence unhashable), byte arrays otherwise provide the same interface and functionality as immutable bytes objects.

The extension module array provides an additional example of a mutable sequence type.

Set types

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 len() 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., 1 and 1.0), only one of them can be contained in a set.

There are currently two intrinsic set types:


These represent a mutable set. They are created by the built-in set() constructor and can be modified afterwards by several methods, such as add().

Frozen sets

These represent an immutable set. They are created by the built-in frozenset() constructor. As a frozenset is immutable and hashable, it can be used again as an element of another set, or as a dictionary key.


These represent finite sets of objects indexed by arbitrary index sets. The subscript notation a[k] selects the item indexed by k from the mapping a; this can be used in expressions and as the target of assignments or del statements. The built-in function len() returns the number of items in a mapping.

There is currently a single intrinsic mapping type:


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’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., 1 and 1.0) then they can be used interchangeably to index the same dictionary entry.

Dictionaries are mutable; they can be created by the {...} notation (see section Dictionary displays).

The extension modules dbm, gdbm, and bsddb provide additional examples of mapping types.

Callable types

These are the types to which the function call operation (see section Calls) can be applied:

User-defined functions

A user-defined function object is created by a function definition (see section Function definitions). It should be called with an argument list containing the same number of items as the function’s formal parameter list.

Special attributes:

Attribute Meaning  
func_doc The function’s documentation string, or None if unavailable Writable
__doc__ Another way of spelling func_doc Writable
func_name The function’s name Writable
__name__ Another way of spelling func_name Writable
__module__ The name of the module the function was defined in, or None if unavailable. Writable
func_defaults A tuple containing default argument values for those arguments that have defaults, or None if no arguments have a default value Writable
func_code The code object representing the compiled function body. Writable
func_globals A reference to the dictionary that holds the function’s global variables — the global namespace of the module in which the function was defined. Read-only
func_dict The namespace supporting arbitrary function attributes. Writable
func_closure None or a tuple of cells that contain bindings for the function’s free variables. Read-only

Most of the attributes labelled “Writable” check the type of the assigned value.

Changed in version 2.4: func_name 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’s definition can be retrieved from its code object; see the description of internal types below.

User-defined methods

A user-defined method object combines a class, a class instance (or None) and any callable object (normally a user-defined function).

Special read-only attributes: im_self is the class instance object, im_func is the function object; im_class is the class of im_self for bound methods or the class that asked for the method for unbound methods; __doc__ is the method’s documentation (same as im_func.__doc__); __name__ is the method name (same as im_func.__name__); __module__ is the name of the module the method was defined in, or None if unavailable.

Changed in version 2.2: im_self used to refer to the class that defined the method.

Changed in version 2.6: For 3.0 forward-compatibility, im_func is also available as __func__, and im_self as __self__.

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 im_self attribute is None 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 im_self attribute is the instance, and the method object is said to be bound. In either case, the new method’s im_class attribute is the class from which the retrieval takes place, and its im_func 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 im_func attribute of the new instance is not the original method object but its im_func attribute.

When a user-defined method object is created by retrieving a class method object from a class or instance, its im_self attribute is the class itself (the same as the im_class attribute), and its im_func attribute is the function object underlying the class method.

When an unbound user-defined method object is called, the underlying function (im_func) is called, with the restriction that the first argument must be an instance of the proper class (im_class) or of a derived class thereof.

When a bound user-defined method object is called, the underlying function (im_func) is called, inserting the class instance (im_self) in front of the argument list. For instance, when C is a class which contains a definition for a function f(), and x is an instance of C, calling x.f(1) is equivalent to calling C.f(x, 1).

When a user-defined method object is derived from a class method object, the “class instance” stored in im_self will actually be the class itself, so that calling either x.f(1) or C.f(1) is equivalent to calling f(C,1) where f 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.

Generator functions

A function or method which uses the yield statement (see section The yield statement) 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’s next() method will cause the function to execute until it provides a value using the yield statement. When the function executes a return statement or falls off the end, a StopIteration exception is raised and the iterator will have reached the end of the set of values to be returned.

Built-in functions

A built-in function object is a wrapper around a C function. Examples of built-in functions are len() and math.sin() (math is a standard built-in module). The number and type of the arguments are determined by the C function. Special read-only attributes: __doc__ is the function’s documentation string, or None if unavailable; __name__ is the function’s name; __self__ is set to None (but see the next item); __module__ is the name of the module the function was defined in or None if unavailable.

Built-in methods

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 alist.append(), assuming alist is a list object. In this case, the special read-only attribute __self__ is set to the object denoted by list.

Class Types
Class types, or “new-style classes,” are callable. These objects normally act as factories for new instances of themselves, but variations are possible for class types that override __new__(). The arguments of the call are passed to __new__() and, in the typical case, to __init__() to initialize the new instance.
Classic Classes

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’s __init__() method if it has one. Any arguments are passed on to the __init__() method. If there is no __init__() method, the class must be called without arguments.

Class instances
Class instances are described below. Class instances are callable only when the class has a __call__() method; x(arguments) is a shorthand for x.__call__(arguments).

Modules are imported by the import statement (see section The import statement). 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., m.x is equivalent to m.__dict__["x"]. A module object does not contain the code object used to initialize the module (since it isn’t needed once the initialization is done).

Attribute assignment updates the module’s namespace dictionary, e.g., m.x = 1 is equivalent to m.__dict__["x"] = 1.

Special read-only attribute: __dict__ is the module’s namespace as a dictionary object.

Predefined (writable) attributes: __name__ is the module’s name; __doc__ is the module’s documentation string, or None if unavailable; __file__ is the pathname of the file from which the module was loaded, if it was loaded from a file. The __file__ 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.


Both class types (new-style classes) and class objects (old-style/classic classes) are typically created by class definitions (see section Class definitions). A class has a namespace implemented by a dictionary object. Class attribute references are translated to lookups in this dictionary, e.g., C.x is translated to C.__dict__["x"] (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 ‘diamond’ 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 http://www.python.org/download/releases/2.3/mro/.

When a class attribute reference (for class C, say) would yield a user-defined function object or an unbound user-defined method object whose associated class is either C or one of its base classes, it is transformed into an unbound user-defined method object whose im_class attribute is C. When it would yield a class method object, it is transformed into a bound user-defined method object whose im_class and im_self attributes are both C. When it would yield a static method object, it is transformed into the object wrapped by the static method object. See section Implementing Descriptors for another way in which attributes retrieved from a class may differ from those actually contained in its __dict__ (note that only new-style classes support descriptors).

Class attribute assignments update the class’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: __name__ is the class name; __module__ is the module name in which the class was defined; __dict__ is the dictionary containing the class’s namespace; __bases__ is a tuple (possibly empty or a singleton) containing the base classes, in the order of their occurrence in the base class list; __doc__ is the class’s documentation string, or None if undefined.

Class instances

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’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 C) 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 im_class attribute is C and whose im_self attribute is the instance. Static method and class method objects are also transformed, as if they had been retrieved from class C; see above under “Classes”. See section Implementing Descriptors for another way in which attributes of a class retrieved via its instances may differ from the objects actually stored in the class’s __dict__. If no class attribute is found, and the object’s class has a __getattr__() method, that is called to satisfy the lookup.

Attribute assignments and deletions update the instance’s dictionary, never a class’s dictionary. If the class has a __setattr__() or __delattr__() 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 Special method names.

Special attributes: __dict__ is the attribute dictionary; __class__ is the instance’s class.


A file object represents an open file. File objects are created by the open() built-in function, and also by os.popen(), os.fdopen(), and the makefile() method of socket objects (and perhaps by other functions or methods provided by extension modules). The objects sys.stdin, sys.stdout and sys.stderr are initialized to file objects corresponding to the interpreter’s standard input, output and error streams. See File Objects for complete documentation of file objects.

Internal types

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.

Code objects

Code objects represent byte-compiled executable Python code, or bytecode. The difference between a code object and a function object is that the function object contains an explicit reference to the function’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: co_name gives the function name; co_argcount is the number of positional arguments (including arguments with default values); co_nlocals is the number of local variables used by the function (including arguments); co_varnames is a tuple containing the names of the local variables (starting with the argument names); co_cellvars is a tuple containing the names of local variables that are referenced by nested functions; co_freevars is a tuple containing the names of free variables; co_code is a string representing the sequence of bytecode instructions; co_consts is a tuple containing the literals used by the bytecode; co_names is a tuple containing the names used by the bytecode; co_filename is the filename from which the code was compiled; co_firstlineno is the first line number of the function; co_lnotab is a string encoding the mapping from bytecode offsets to line numbers (for details see the source code of the interpreter); co_stacksize is the required stack size (including local variables); co_flags is an integer encoding a number of flags for the interpreter.

The following flag bits are defined for co_flags: bit 0x04 is set if the function uses the *arguments syntax to accept an arbitrary number of positional arguments; bit 0x08 is set if the function uses the **keywords syntax to accept arbitrary keyword arguments; bit 0x20 is set if the function is a generator.

Future feature declarations (from __future__ import division) also use bits in co_flags to indicate whether a code object was compiled with a particular feature enabled: bit 0x2000 is set if the function was compiled with future division enabled; bits 0x10 and 0x1000 were used in earlier versions of Python.

Other bits in co_flags are reserved for internal use.

If a code object represents a function, the first item in co_consts is the documentation string of the function, or None if undefined.

Frame objects

Frame objects represent execution frames. They may occur in traceback objects (see below).

Special read-only attributes: f_back is to the previous stack frame (towards the caller), or None if this is the bottom stack frame; f_code is the code object being executed in this frame; f_locals is the dictionary used to look up local variables; f_globals is used for global variables; f_builtins is used for built-in (intrinsic) names; f_restricted is a flag indicating whether the function is executing in restricted execution mode; f_lasti gives the precise instruction (this is an index into the bytecode string of the code object).

Special writable attributes: f_trace, if not None, is a function called at the start of each source code line (this is used by the debugger); f_exc_type, f_exc_value, f_exc_traceback 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); f_lineno is the current line number of the frame — 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.

Traceback objects

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 The try statement.) It is accessible as sys.exc_traceback, and also as the third item of the tuple returned by sys.exc_info(). 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 sys.last_traceback.

Special read-only attributes: tb_next is the next level in the stack trace (towards the frame where the exception occurred), or None if there is no next level; tb_frame points to the execution frame of the current level; tb_lineno gives the line number where the exception occurred; tb_lasti 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 try statement with no matching except clause or with a finally clause.

Slice objects

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., a[i:j:step], a[i:j, k:l], or a[..., i:j]. They are also created by the built-in slice() function.

Special read-only attributes: start is the lower bound; stop is the upper bound; step is the step value; each is None if omitted. These attributes can have any type.

Slice objects support one method:

slice.indices(self, length)

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.

Static method objects
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 staticmethod() constructor.
Class method objects
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 “User-defined methods”. Class method objects are created by the built-in classmethod() constructor.

3.3. New-style and classic classes

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 x.__class__ designates the class of x, but type(x) is always <type 'instance'>. This reflects the fact that all old-style instances, independently of their class, are implemented with a single built-in type, called instance.

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 type(x) is typically the same as x.__class__ (although this is not guaranteed - a new-style class instance is permitted to override the value returned for x.__class__).

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 “descriptors”, 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 “top-level type” object 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 type() returns. Some of these changes are fundamental to the new object model, like the way special methods are invoked. Others are “fixes” 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’s class mechanics, it may still be lacking in some areas when it comes to its coverage of new-style classes. Please see http://www.python.org/doc/newstyle/ for sources of additional information.

Old-style classes are removed in Python 3.0, leaving only the semantics of new-style classes.

3.4. Special method names

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’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 __getitem__(), and x is an instance of this class, then x[i] is roughly equivalent to x.__getitem__(i) for old-style classes and type(x).__getitem__(x, i) for new-style classes. Except where mentioned, attempts to execute an operation raise an exception when no appropriate method is defined (typically AttributeError or TypeError).

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 NodeList interface in the W3C’s Document Object Model.)

3.4.1. Basic customization

object.__new__(cls[, ...])

Called to create a new instance of class cls. __new__() 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 __new__() should be the new object instance (usually an instance of cls).

Typical implementations create a new instance of the class by invoking the superclass’s __new__() method using super(currentclass, cls).__new__(cls[, ...]) with appropriate arguments and then modifying the newly-created instance as necessary before returning it.

If __new__() returns an instance of cls, then the new instance’s __init__() method will be invoked like __init__(self[, ...]), where self is the new instance and the remaining arguments are the same as were passed to __new__().

If __new__() does not return an instance of cls, then the new instance’s __init__() method will not be invoked.

__new__() 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.

object.__init__(self[, ...])

Called when the instance is created. The arguments are those passed to the class constructor expression. If a base class has an __init__() method, the derived class’s __init__() method, if any, must explicitly call it to ensure proper initialization of the base class part of the instance; for example: BaseClass.__init__(self, [args...]). As a special constraint on constructors, no value may be returned; doing so will cause a TypeError to be raised at runtime.


Called when the instance is about to be destroyed. This is also called a destructor. If a base class has a __del__() method, the derived class’s __del__() 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 __del__() 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 __del__() methods are called for objects that still exist when the interpreter exits.


del x doesn’t directly call x.__del__() — the former decrements the reference count for x by one, and the latter is only called when x‘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 sys.exc_traceback 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 sys.last_traceback 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 None in sys.exc_traceback or sys.last_traceback. Circular references which are garbage are detected when the option cycle detector is enabled (it’s on by default), but can only be cleaned up if there are no Python-level __del__() methods involved. Refer to the documentation for the gc module for more information about how __del__() methods are handled by the cycle detector, particularly the description of the garbage value.


Due to the precarious circumstances under which __del__() methods are invoked, exceptions that occur during their execution are ignored, and a warning is printed to sys.stderr instead. Also, when __del__() is invoked in response to a module being deleted (e.g., when execution of the program is done), other globals referenced by the __del__() method may already have been deleted or in the process of being torn down (e.g. the import machinery shutting down). For this reason, __del__() 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 __del__() method is called.


Called by the repr() built-in function and by string conversions (reverse quotes) to compute the “official” 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 <...some useful description...> should be returned. The return value must be a string object. If a class defines __repr__() but not __str__(), then __repr__() is also used when an “informal” 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.


Called by the str() built-in function and by the print statement to compute the “informal” string representation of an object. This differs from __repr__() 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.

object.__lt__(self, other)
object.__le__(self, other)
object.__eq__(self, other)
object.__ne__(self, other)
object.__gt__(self, other)
object.__ge__(self, other)

New in version 2.1.

These are the so-called “rich comparison” methods, and are called for comparison operators in preference to __cmp__() below. The correspondence between operator symbols and method names is as follows: x<y calls x.__lt__(y), x<=y calls x.__le__(y), x==y calls x.__eq__(y), x!=y and x<>y call x.__ne__(y), x>y calls x.__gt__(y), and x>=y calls x.__ge__(y).

A rich comparison method may return the singleton NotImplemented if it does not implement the operation for a given pair of arguments. By convention, False and True 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 if statement), Python will call bool() on the value to determine if the result is true or false.

There are no implied relationships among the comparison operators. The truth of x==y does not imply that x!=y is false. Accordingly, when defining __eq__(), one should also define __ne__() so that the operators will behave as expected. See the paragraph on __hash__() for some important notes on creating hashable 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, __lt__() and __gt__() are each other’s reflection, __le__() and __ge__() are each other’s reflection, and __eq__() and __ne__() are their own reflection.

Arguments to rich comparison methods are never coerced.

To automatically generate ordering operations from a single root operation, see the Total Ordering recipe in the ASPN cookbook.

object.__cmp__(self, other)

Called by comparison operations if rich comparison (see above) is not defined. Should return a negative integer if self < other, zero if self == other, a positive integer if self > other. If no __cmp__(), __eq__() or __ne__() operation is defined, class instances are compared by object identity (“address”). See also the description of __hash__() for some important notes on creating hashable objects which support custom comparison operations and are usable as dictionary keys. (Note: the restriction that exceptions are not propagated by __cmp__() has been removed since Python 1.5.)

object.__rcmp__(self, other)

Changed in version 2.1: No longer supported.


Called by built-in function hash() and for operations on members of hashed collections including set, frozenset, and dict. __hash__() 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 __cmp__() or __eq__() method it should not define a __hash__() operation either; if it defines __cmp__() or __eq__() but not __hash__(), its instances will not be usable in hashed collections. If a class defines mutable objects and implements a __cmp__() or __eq__() method, it should not implement __hash__(), since hashable collection implementations require that a object’s hash value is immutable (if the object’s hash value changes, it will be in the wrong hash bucket).

User-defined classes have __cmp__() and __hash__() methods by default; with them, all objects compare unequal (except with themselves) and x.__hash__() returns id(x).

Classes which inherit a __hash__() method from a parent class but change the meaning of __cmp__() or __eq__() 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 __hash__ = None in the class definition. Doing so means that not only will instances of the class raise an appropriate TypeError when a program attempts to retrieve their hash value, but they will also be correctly identified as unhashable when checking isinstance(obj, collections.Hashable) (unlike classes which define their own __hash__() to explicitly raise TypeError).

Changed in version 2.5: __hash__() 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: __hash__ may now be set to None to explicitly flag instances of a class as unhashable.


Called to implement truth value testing and the built-in operation bool(); should return False or True, or their integer equivalents 0 or 1. When this method is not defined, __len__() is called, if it is defined, and the object is considered true if its result is nonzero. If a class defines neither __len__() nor __nonzero__(), all its instances are considered true.


Called to implement unicode() 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.

3.4.2. Customizing attribute access

The following methods can be defined to customize the meaning of attribute access (use of, assignment to, or deletion of x.name) for class instances.

object.__getattr__(self, name)

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 self). name is the attribute name. This method should return the (computed) attribute value or raise an AttributeError exception.

Note that if the attribute is found through the normal mechanism, __getattr__() is not called. (This is an intentional asymmetry between __getattr__() and __setattr__().) This is done both for efficiency reasons and because otherwise __getattr__() 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 __getattribute__() method below for a way to actually get total control in new-style classes.

object.__setattr__(self, name, value)

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 __setattr__() wants to assign to an instance attribute, it should not simply execute self.name = value — this would cause a recursive call to itself. Instead, it should insert the value in the dictionary of instance attributes, e.g., self.__dict__[name] = value. For new-style classes, rather than accessing the instance dictionary, it should call the base class method with the same name, for example, object.__setattr__(self, name, value).

object.__delattr__(self, name)
Like __setattr__() but for attribute deletion instead of assignment. This should only be implemented if del obj.name is meaningful for the object. More attribute access for new-style classes

The following methods only apply to new-style classes.

object.__getattribute__(self, name)

Called unconditionally to implement attribute accesses for instances of the class. If the class also defines __getattr__(), the latter will not be called unless __getattribute__() either calls it explicitly or raises an AttributeError. This method should return the (computed) attribute value or raise an AttributeError 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, object.__getattribute__(self, name).


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 Special method lookup for new-style classes. Implementing Descriptors

The following methods only apply when an instance of the class containing the method (a so-called descriptor class) appears in the class dictionary of another new-style class, known as the owner class. In the examples below, “the attribute” refers to the attribute whose name is the key of the property in the owner class’ __dict__. Descriptors can only be implemented as new-style classes themselves.

object.__get__(self, instance, owner)
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 None when the attribute is accessed through the owner. This method should return the (computed) attribute value or raise an AttributeError exception.
object.__set__(self, instance, value)
Called to set the attribute on an instance instance of the owner class to a new value, value.
object.__delete__(self, instance)
Called to delete the attribute on an instance instance of the owner class. Invoking Descriptors

In general, a descriptor is an object attribute with “binding behavior”, one whose attribute access has been overridden by methods in the descriptor protocol: __get__(), __set__(), and __delete__(). 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’s dictionary. For instance, a.x has a lookup chain starting with a.__dict__['x'], then type(a).__dict__['x'], and continuing through the base classes of type(a) 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 object() or type()).

The starting point for descriptor invocation is a binding, a.x. How the arguments are assembled depends on a:

Direct Call
The simplest and least common call is when user code directly invokes a descriptor method: x.__get__(a).
Instance Binding
If binding to a new-style object instance, a.x is transformed into the call: type(a).__dict__['x'].__get__(a, type(a)).
Class Binding
If binding to a new-style class, A.x is transformed into the call: A.__dict__['x'].__get__(None, A).
Super Binding
If a is an instance of super, then the binding super(B, obj).m() searches obj.__class__.__mro__ for the base class A immediately preceding B and then invokes the descriptor with the call: A.__dict__['m'].__get__(obj, A).

For instance bindings, the precedence of descriptor invocation depends on the which descriptor methods are defined. Normally, data descriptors define both __get__() and __set__(), while non-data descriptors have just the __get__() method. Data descriptors always override a redefinition in an instance dictionary. In contrast, non-data descriptors can be overridden by instances. [2]

Python methods (including staticmethod() and classmethod()) 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 property() function is implemented as a data descriptor. Accordingly, instances cannot override the behavior of a property. __slots__

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.


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.

Notes on using __slots__

  • 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.

  • 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 AttributeError. If dynamic assignment of new variables is desired, then add '__dict__' to the sequence of strings in the __slots__ declaration.

    Changed in version 2.3: Previously, adding '__dict__' to the __slots__ declaration would not enable the assignment of new attributes not specifically listed in the sequence of instance variable names.

  • 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 '__weakref__' to the sequence of strings in the __slots__ declaration.

    Changed in version 2.3: Previously, adding '__weakref__' to the __slots__ declaration would not enable support for weak references.

  • __slots__ are implemented at the class level by creating descriptors (Implementing Descriptors) 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.

  • 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).

  • 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.

  • Nonempty __slots__ does not work for classes derived from “variable-length” built-in types such as long, str and tuple.

  • 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.

  • __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__.

3.4.3. Customizing class creation

By default, new-style classes are constructed using type(). A class definition is read into a separate namespace and the value of class name is bound to the result of type(name, bases, dict).

When the class definition is read, if __metaclass__ is defined then the callable assigned to it will be called instead of type(). This allows classes or functions to be written which monitor or alter the class creation process:

  • Modifying the class dictionary prior to the class being created.
  • Returning an instance of another class – essentially performing the role of a factory function.

These steps will have to be performed in the metaclass’s __new__() method – type.__new__() 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:

class metacls(type):
    def __new__(mcs, name, bases, dict):
        dict['foo'] = 'metacls was here'
        return type.__new__(mcs, name, bases, dict)

You can of course also override other class methods (or add new methods); for example defining a custom __call__() method in the metaclass allows custom behavior when the class is called, e.g. not always creating a new instance.


This variable can be any callable accepting arguments for name, bases, and dict. Upon class creation, the callable is used instead of the built-in type().

New in version 2.2.

The appropriate metaclass is determined by the following precedence rules:

  • If dict['__metaclass__'] exists, it is used.
  • 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).
  • Otherwise, if a global variable named __metaclass__ exists, it is used.
  • Otherwise, the old-style, classic metaclass (types.ClassType) is used.

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.

3.4.4. Emulating callable objects

object.__call__(self[, args...])

Called when the instance is “called” as a function; if this method is defined, x(arg1, arg2, ...) is a shorthand for x.__call__(arg1, arg2, ...).

3.4.5. Emulating container types

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 0 <= k < N where N is the length of the sequence, or slice objects, which define a range of items. (For backwards compatibility, the method __getslice__() (see below) can also be defined to handle simple, but not extended slices.) It is also recommended that mappings provide the methods keys(), values(), items(), has_key(), get(), clear(), setdefault(), iterkeys(), itervalues(), iteritems(), pop(), popitem(), copy(), and update() behaving similar to those for Python’s standard dictionary objects. The UserDict module provides a DictMixin class to help create those methods from a base set of __getitem__(), __setitem__(), __delitem__(), and keys(). Mutable sequences should provide methods append(), count(), index(), extend(), insert(), pop(), remove(), reverse() and sort(), like Python standard list objects. Finally, sequence types should implement addition (meaning concatenation) and multiplication (meaning repetition) by defining the methods __add__(), __radd__(), __iadd__(), __mul__(), __rmul__() and __imul__() described below; they should not define __coerce__() or other numerical operators. It is recommended that both mappings and sequences implement the __contains__() method to allow efficient use of the in operator; for mappings, in should be equivalent of has_key(); for sequences, it should search through the values. It is further recommended that both mappings and sequences implement the __iter__() method to allow efficient iteration through the container; for mappings, __iter__() should be the same as iterkeys(); for sequences, it should iterate through the values.


Called to implement the built-in function len(). Should return the length of the object, an integer >= 0. Also, an object that doesn’t define a __nonzero__() method and whose __len__() method returns zero is considered to be false in a Boolean context.

object.__getitem__(self, key)

Called to implement evaluation of self[key]. 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 __getitem__() method. If key is of an inappropriate type, TypeError may be raised; if of a value outside the set of indexes for the sequence (after any special interpretation of negative values), IndexError should be raised. For mapping types, if key is missing (not in the container), KeyError should be raised.


for loops expect that an IndexError will be raised for illegal indexes to allow proper detection of the end of the sequence.

object.__setitem__(self, key, value)
Called to implement assignment to self[key]. Same note as for __getitem__(). 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 __getitem__() method.
object.__delitem__(self, key)
Called to implement deletion of self[key]. Same note as for __getitem__(). 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 __getitem__() method.

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 iterkeys().

Iterator objects also need to implement this method; they are required to return themselves. For more information on iterator objects, see Iterator Types.


Called (if present) by the reversed() 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 __reversed__() method is not provided, the reversed() built-in will fall back to using the sequence protocol (__len__() and __getitem__()). Objects that support the sequence protocol should only provide __reversed__() if they can provide an implementation that is more efficient than the one provided by reversed().

New in version 2.6.

The membership test operators (in and not in) 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.

object.__contains__(self, item)

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’t define __contains__(), the membership test first tries iteration via __iter__(), then the old sequence iteration protocol via __getitem__(), see this section in the language reference.

3.4.6. Additional methods for emulation of sequence types

The following optional methods can be defined to further emulate sequence objects. Immutable sequences methods should at most only define __getslice__(); mutable sequences might define all three methods.

object.__getslice__(self, i, j)

Deprecated since version 2.0: Support slice objects as parameters to the __getitem__() method. (However, built-in types in CPython currently still implement __getslice__(). Therefore, you have to override it in derived classes when implementing slicing.)

Called to implement evaluation of self[i:j]. 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 sys.maxint, 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 __len__() method, an AttributeError 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 __getslice__() is found, a slice object is created instead, and passed to __getitem__() instead.

object.__setslice__(self, i, j, sequence)

Called to implement assignment to self[i:j]. Same notes for i and j as for __getslice__().

This method is deprecated. If no __setslice__() is found, or for extended slicing of the form self[i:j:k], a slice object is created, and passed to __setitem__(), instead of __setslice__() being called.

object.__delslice__(self, i, j)
Called to implement deletion of self[i:j]. Same notes for i and j as for __getslice__(). This method is deprecated. If no __delslice__() is found, or for extended slicing of the form self[i:j:k], a slice object is created, and passed to __delitem__(), instead of __delslice__() being called.

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, __getitem__(), __setitem__() or __delitem__() 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 __getitem__(), __setitem__() and __delitem__() support slice objects as arguments):

class MyClass:
    def __getitem__(self, index):
    def __setitem__(self, index, value):
    def __delitem__(self, index):

    if sys.version_info < (2, 0):
        # They won'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):]

Note the calls to max(); these are necessary because of the handling of negative indices before the __*slice__() methods are called. When negative indexes are used, the __*item__() methods receive them as provided, but the __*slice__() methods get a “cooked” 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 __*item__() 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 __*item__() methods. Calling max(0, i) conveniently returns the proper value.

3.4.7. Emulating numeric types

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.

object.__add__(self, other)
object.__sub__(self, other)
object.__mul__(self, other)
object.__floordiv__(self, other)
object.__mod__(self, other)
object.__divmod__(self, other)
object.__pow__(self, other[, modulo])
object.__lshift__(self, other)
object.__rshift__(self, other)
object.__and__(self, other)
object.__xor__(self, other)
object.__or__(self, other)

These methods are called to implement the binary arithmetic operations (+, -, *, //, %, divmod(), pow(), **, <<, >>, &, ^, |). For instance, to evaluate the expression x + y, where x is an instance of a class that has an __add__() method, x.__add__(y) is called. The __divmod__() method should be the equivalent to using __floordiv__() and __mod__(); it should not be related to __truediv__() (described below). Note that __pow__() should be defined to accept an optional third argument if the ternary version of the built-in pow() function is to be supported.

If one of those methods does not support the operation with the supplied arguments, it should return NotImplemented.

object.__div__(self, other)
object.__truediv__(self, other)
The division operator (/) is implemented by these methods. The __truediv__() method is used when __future__.division is in effect, otherwise __div__() is used. If only one of these two methods is defined, the object will not support division in the alternate context; TypeError will be raised instead.
object.__radd__(self, other)
object.__rsub__(self, other)
object.__rmul__(self, other)
object.__rdiv__(self, other)
object.__rtruediv__(self, other)
object.__rfloordiv__(self, other)
object.__rmod__(self, other)
object.__rdivmod__(self, other)
object.__rpow__(self, other)
object.__rlshift__(self, other)
object.__rrshift__(self, other)
object.__rand__(self, other)
object.__rxor__(self, other)
object.__ror__(self, other)

These methods are called to implement the binary arithmetic operations (+, -, *, /, %, divmod(), pow(), **, <<, >>, &, ^, |) 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. [3] For instance, to evaluate the expression x - y, where y is an instance of a class that has an __rsub__() method, y.__rsub__(x) is called if x.__sub__(y) returns NotImplemented.

Note that ternary pow() will not try calling __rpow__() (the coercion rules would become too complicated).


If the right operand’s type is a subclass of the left operand’s type and that subclass provides the reflected method for the operation, this method will be called before the left operand’s non-reflected method. This behavior allows subclasses to override their ancestors’ operations.

object.__iadd__(self, other)
object.__isub__(self, other)
object.__imul__(self, other)
object.__idiv__(self, other)
object.__itruediv__(self, other)
object.__ifloordiv__(self, other)
object.__imod__(self, other)
object.__ipow__(self, other[, modulo])
object.__ilshift__(self, other)
object.__irshift__(self, other)
object.__iand__(self, other)
object.__ixor__(self, other)
object.__ior__(self, other)
These methods are called to implement the augmented arithmetic assignments (+=, -=, *=, /=, //=, %=, **=, <<=, >>=, &=, ^=, |=). 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 x += y, where x is an instance of a class that has an __iadd__() method, x.__iadd__(y) is called. If x is an instance of a class that does not define a __iadd__() method, x.__add__(y) and y.__radd__(x) are considered, as with the evaluation of x + y.

Called to implement the unary arithmetic operations (-, +, abs() and ~).


Called to implement the built-in functions complex(), int(), long(), and float(). Should return a value of the appropriate type.


Called to implement the built-in functions oct() and hex(). Should return a string value.


Called to implement operator.index(). Also called whenever Python needs an integer object (such as in slicing). Must return an integer (int or long).

New in version 2.5.

object.__coerce__(self, other)
Called to implement “mixed-mode” numeric arithmetic. Should either return a 2-tuple containing self and other converted to a common numeric type, or None if conversion is impossible. When the common type would be the type of other, it is sufficient to return None, 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 NotImplemented is equivalent to returning None.

3.4.8. Coercion rules

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 3.0, coercion will not be supported.

  • 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.

  • It is no longer recommended to define a coercion operation. Mixed-mode operations on types that don’t define coercion pass the original arguments to the operation.

  • New-style classes (those derived from object) never invoke the __coerce__() method in response to a binary operator; the only time __coerce__() is invoked is when the built-in function coerce() is called.

  • For most intents and purposes, an operator that returns NotImplemented is treated the same as one that is not implemented at all.

  • Below, __op__() and __rop__() are used to signify the generic method names corresponding to an operator; __iop__() is used for the corresponding in-place operator. For example, for the operator ‘+‘, __add__() and __radd__() are used for the left and right variant of the binary operator, and __iadd__() for the in-place variant.

  • For objects x and y, first x.__op__(y) is tried. If this is not implemented or returns NotImplemented, y.__rop__(x) is tried. If this is also not implemented or returns NotImplemented, a TypeError exception is raised. But see the following exception:

  • 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’s __rop__() method, the right operand’s __rop__() method is tried before the left operand’s __op__() method.

    This is done so that a subclass can completely override binary operators. Otherwise, the left operand’s __op__() 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.

  • When either operand type defines a coercion, this coercion is called before that type’s __op__() or __rop__() 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.

  • When an in-place operator (like ‘+=‘) is used, if the left operand implements __iop__(), it is invoked without any coercion. When the operation falls back to __op__() and/or __rop__(), the normal coercion rules apply.

  • In x + y, if x is a sequence that implements sequence concatenation, sequence concatenation is invoked.

  • In x * y, if one operator is a sequence that implements sequence repetition, and the other is an integer (int or long), sequence repetition is invoked.

  • Rich comparisons (implemented by methods __eq__() and so on) never use coercion. Three-way comparison (implemented by __cmp__()) does use coercion under the same conditions as other binary operations use it.

  • In the current implementation, the built-in numeric types int, long and float do not use coercion; the type complex however does use coercion for binary operators and rich comparisons, despite the above rules. The difference can become apparent when subclassing these types. Over time, the type complex may be fixed to avoid coercion. All these types implement a __coerce__() method, for use by the built-in coerce() function.

3.4.9. With Statement Context Managers

New in version 2.5.

A context manager is an object that defines the runtime context to be established when executing a with 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 with statement (described in section The with statement), 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 Context Manager Types.

Enter the runtime context related to this object. The with statement will bind this method’s return value to the target(s) specified in the as clause of the statement, if any.
object.__exit__(self, exc_type, exc_value, traceback)

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 None.

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 __exit__() methods should not reraise the passed-in exception; this is the caller’s responsibility.

See also

PEP 0343 - The “with” statement
The specification, background, and examples for the Python with statement.

3.4.10. Special method lookup for old-style classes

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 x.__getitem__(i) or implicitly as in x[i].

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:

>>> class C:
...     pass
>>> c1 = C()
>>> c2 = C()
>>> c1.__len__ = lambda: 5
>>> c2.__len__ = lambda: 9
>>> len(c1)
>>> len(c2)

3.4.11. Special method lookup for new-style classes

For new-style classes, implicit invocations of special methods are only guaranteed to work correctly if defined on an object’s type, not in the object’s instance dictionary. That behaviour is the reason why the following code raises an exception (unlike the equivalent example with old-style classes):

>>> class C(object):
...     pass
>>> c = C()
>>> c.__len__ = lambda: 5
>>> len(c)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: object of type 'C' has no len()

The rationale behind this behaviour lies with a number of special methods such as __hash__() and __repr__() 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:

>>> 1 .__hash__() == hash(1)
>>> int.__hash__() == hash(int)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: descriptor '__hash__' of 'int' object needs an argument

Incorrectly attempting to invoke an unbound method of a class in this way is sometimes referred to as ‘metaclass confusion’, and is avoided by bypassing the instance when looking up special methods:

>>> type(1).__hash__(1) == hash(1)
>>> type(int).__hash__(int) == hash(int)

In addition to bypassing any instance attributes in the interest of correctness, implicit special method lookup generally also bypasses the __getattribute__() method even of the object’s metaclass:

>>> class Meta(type):
...    def __getattribute__(*args):
...       print "Metaclass getattribute invoked"
...       return type.__getattribute__(*args)
>>> class C(object):
...     __metaclass__ = Meta
...     def __len__(self):
...         return 10
...     def __getattribute__(*args):
...         print "Class getattribute invoked"
...         return object.__getattribute__(*args)
>>> c = C()
>>> c.__len__()                 # Explicit lookup via instance
Class getattribute invoked
>>> type(c).__len__(c)          # Explicit lookup via type
Metaclass getattribute invoked
>>> len(c)                      # Implicit lookup

Bypassing the __getattribute__() 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).


[1]It is possible in some cases to change an object’s type, under certain controlled conditions. It generally isn’t a good idea though, since it can lead to some very strange behaviour if it is handled incorrectly.
[2]A descriptor can define any combination of __get__(), __set__() and __delete__(). If it does not define __get__(), then accessing the attribute even on an instance will return the descriptor object itself. If the descriptor defines __set__() and/or __delete__(), it is a data descriptor; if it defines neither, it is a non-data descriptor.
[3]For operands of the same type, it is assumed that if the non-reflected method (such as __add__()) fails the operation is not supported, which is why the reflected method is not called.