Doc/lib/liboperator.tex

\section{\module{operator} ---
         Standard operators as functions.}
\declaremodule{builtin}{operator}
\sectionauthor{Skip Montanaro}{skip@automatrix.com}

\modulesynopsis{All Python's standard operators as built-in functions.}


The \module{operator} module exports a set of functions implemented in C
corresponding to the intrinsic operators of Python.  For example,
\code{operator.add(x, y)} is equivalent to the expression \code{x+y}.  The
function names are those used for special class methods; variants without
leading and trailing \samp{__} are also provided for convenience.

The functions fall into categories that perform object comparisons,
logical operations, mathematical operations, sequence operations, and
abstract type tests.

The object comparison functions are useful for all objects, and are
named after the rich comparison operators they support:

\begin{funcdesc}{lt}{a, b}
\funcline{le}{a, b}
\funcline{eq}{a, b}
\funcline{ne}{a, b}
\funcline{ge}{a, b}
\funcline{gt}{a, b}
\funcline{__lt__}{a, b}
\funcline{__le__}{a, b}
\funcline{__eq__}{a, b}
\funcline{__ne__}{a, b}
\funcline{__ge__}{a, b}
\funcline{__gt__}{a, b}
Perform ``rich comparisons'' between \var{a} and \var{b}. Specifically,
\code{lt(\var{a}, \var{b})} is equivalent to \code{\var{a} < \var{b}},
\code{le(\var{a}, \var{b})} is equivalent to \code{\var{a} <= \var{b}},
\code{eq(\var{a}, \var{b})} is equivalent to \code{\var{a} == \var{b}},
\code{ne(\var{a}, \var{b})} is equivalent to \code{\var{a} != \var{b}},
\code{gt(\var{a}, \var{b})} is equivalent to \code{\var{a} > \var{b}}
and
\code{ge(\var{a}, \var{b})} is equivalent to \code{\var{a} >= \var{b}}.
Note that unlike the built-in \function{cmp()}, these functions can
return any value, which may or may not be interpretable as a Boolean
value.  See the \citetitle[../ref/ref.html]{Python Reference Manual}
for more information about rich comparisons.
\versionadded{2.2}
\end{funcdesc}


The logical operations are also generally applicable to all objects,
and support truth tests, identity tests, and boolean operations:

\begin{funcdesc}{not_}{o}
\funcline{__not__}{o}
Return the outcome of \keyword{not} \var{o}.  (Note that there is no
\method{__not__()} method for object instances; only the interpreter
core defines this operation.  The result is affected by the
\method{__bool__()} and \method{__len__()} methods.)
\end{funcdesc}

\begin{funcdesc}{truth}{o}
Return \constant{True} if \var{o} is true, and \constant{False}
otherwise.  This is equivalent to using the \class{bool}
constructor.
\end{funcdesc}

\begin{funcdesc}{is_}{a, b}
Return \code{\var{a} is \var{b}}.  Tests object identity.
\versionadded{2.3}
\end{funcdesc}

\begin{funcdesc}{is_not}{a, b}
Return \code{\var{a} is not \var{b}}.  Tests object identity.
\versionadded{2.3}
\end{funcdesc}


The mathematical and bitwise operations are the most numerous:

\begin{funcdesc}{abs}{o}
\funcline{__abs__}{o}
Return the absolute value of \var{o}.
\end{funcdesc}

\begin{funcdesc}{add}{a, b}
\funcline{__add__}{a, b}
Return \var{a} \code{+} \var{b}, for \var{a} and \var{b} numbers.
\end{funcdesc}

\begin{funcdesc}{and_}{a, b}
\funcline{__and__}{a, b}
Return the bitwise and of \var{a} and \var{b}.
\end{funcdesc}

\begin{funcdesc}{div}{a, b}
\funcline{__div__}{a, b}
Return \var{a} \code{/} \var{b} when \code{__future__.division} is not
in effect.  This is also known as ``classic'' division.
\end{funcdesc}

\begin{funcdesc}{floordiv}{a, b}
\funcline{__floordiv__}{a, b}
Return \var{a} \code{//} \var{b}.
\versionadded{2.2}
\end{funcdesc}

\begin{funcdesc}{inv}{o}
\funcline{invert}{o}
\funcline{__inv__}{o}
\funcline{__invert__}{o}
Return the bitwise inverse of the number \var{o}.  This is equivalent
to \code{\textasciitilde}\var{o}.  The names \function{invert()} and
\function{__invert__()} were added in Python 2.0.
\end{funcdesc}

\begin{funcdesc}{lshift}{a, b}
\funcline{__lshift__}{a, b}
Return \var{a} shifted left by \var{b}.
\end{funcdesc}

\begin{funcdesc}{mod}{a, b}
\funcline{__mod__}{a, b}
Return \var{a} \code{\%} \var{b}.
\end{funcdesc}

\begin{funcdesc}{mul}{a, b}
\funcline{__mul__}{a, b}
Return \var{a} \code{*} \var{b}, for \var{a} and \var{b} numbers.
\end{funcdesc}

\begin{funcdesc}{neg}{o}
\funcline{__neg__}{o}
Return \var{o} negated.
\end{funcdesc}

\begin{funcdesc}{or_}{a, b}
\funcline{__or__}{a, b}
Return the bitwise or of \var{a} and \var{b}.
\end{funcdesc}

\begin{funcdesc}{pos}{o}
\funcline{__pos__}{o}
Return \var{o} positive.
\end{funcdesc}

\begin{funcdesc}{pow}{a, b}
\funcline{__pow__}{a, b}
Return \var{a} \code{**} \var{b}, for \var{a} and \var{b} numbers.
\versionadded{2.3}
\end{funcdesc}

\begin{funcdesc}{rshift}{a, b}
\funcline{__rshift__}{a, b}
Return \var{a} shifted right by \var{b}.
\end{funcdesc}

\begin{funcdesc}{sub}{a, b}
\funcline{__sub__}{a, b}
Return \var{a} \code{-} \var{b}.
\end{funcdesc}

\begin{funcdesc}{truediv}{a, b}
\funcline{__truediv__}{a, b}
Return \var{a} \code{/} \var{b} when \code{__future__.division} is in
effect.  This is also known as ``true'' division.
\versionadded{2.2}
\end{funcdesc}

\begin{funcdesc}{xor}{a, b}
\funcline{__xor__}{a, b}
Return the bitwise exclusive or of \var{a} and \var{b}.
\end{funcdesc}

\begin{funcdesc}{index}{a}
\funcline{__index__}{a}
Return \var{a} converted to an integer.  Equivalent to \var{a}\code{.__index__()}.
\versionadded{2.5}
\end{funcdesc}

Operations which work with sequences include:

\begin{funcdesc}{concat}{a, b}
\funcline{__concat__}{a, b}
Return \var{a} \code{+} \var{b} for \var{a} and \var{b} sequences.
\end{funcdesc}

\begin{funcdesc}{contains}{a, b}
\funcline{__contains__}{a, b}
Return the outcome of the test \var{b} \code{in} \var{a}.
Note the reversed operands.  The name \function{__contains__()} was
added in Python 2.0.
\end{funcdesc}

\begin{funcdesc}{countOf}{a, b}
Return the number of occurrences of \var{b} in \var{a}.
\end{funcdesc}

\begin{funcdesc}{delitem}{a, b}
\funcline{__delitem__}{a, b}
Remove the value of \var{a} at index \var{b}.
\end{funcdesc}

\begin{funcdesc}{delslice}{a, b, c}
\funcline{__delslice__}{a, b, c}
Delete the slice of \var{a} from index \var{b} to index \var{c}\code{-1}.
\end{funcdesc}

\begin{funcdesc}{getitem}{a, b}
\funcline{__getitem__}{a, b}
Return the value of \var{a} at index \var{b}.
\end{funcdesc}

\begin{funcdesc}{getslice}{a, b, c}
\funcline{__getslice__}{a, b, c}
Return the slice of \var{a} from index \var{b} to index \var{c}\code{-1}.
\end{funcdesc}

\begin{funcdesc}{indexOf}{a, b}
Return the index of the first of occurrence of \var{b} in \var{a}.
\end{funcdesc}

\begin{funcdesc}{repeat}{a, b}
\funcline{__repeat__}{a, b}
Return \var{a} \code{*} \var{b} where \var{a} is a sequence and
\var{b} is an integer.
\end{funcdesc}

\begin{funcdesc}{sequenceIncludes}{\unspecified}
\deprecated{2.0}{Use \function{contains()} instead.}
Alias for \function{contains()}.
\end{funcdesc}

\begin{funcdesc}{setitem}{a, b, c}
\funcline{__setitem__}{a, b, c}
Set the value of \var{a} at index \var{b} to \var{c}.
\end{funcdesc}

\begin{funcdesc}{setslice}{a, b, c, v}
\funcline{__setslice__}{a, b, c, v}
Set the slice of \var{a} from index \var{b} to index \var{c}\code{-1} to the
sequence \var{v}.
\end{funcdesc}


Many operations have an ``in-place'' version.  The following functions
provide a more primitive access to in-place operators than the usual
syntax does; for example, the statement \code{x += y} is equivalent to
\code{x = operator.iadd(x, y)}.  Another way to put it is to say that
\code{z = operator.iadd(x, y)} is equivalent to the compound statement
\code{z = x; z += y}.

\begin{funcdesc}{iadd}{a, b}
\funcline{__iadd__}{a, b}
\code{a = iadd(a, b)} is equivalent to \code{a += b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{iand}{a, b}
\funcline{__iand__}{a, b}
\code{a = iand(a, b)} is equivalent to \code{a \&= b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{iconcat}{a, b}
\funcline{__iconcat__}{a, b}
\code{a = iconcat(a, b)} is equivalent to \code{a += b} for \var{a}
and \var{b} sequences.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{idiv}{a, b}
\funcline{__idiv__}{a, b}
\code{a = idiv(a, b)} is equivalent to \code{a /= b} when
\code{__future__.division} is not in effect.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{ifloordiv}{a, b}
\funcline{__ifloordiv__}{a, b}
\code{a = ifloordiv(a, b)} is equivalent to \code{a //= b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{ilshift}{a, b}
\funcline{__ilshift__}{a, b}
\code{a = ilshift(a, b)} is equivalent to \code{a <}\code{<= b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{imod}{a, b}
\funcline{__imod__}{a, b}
\code{a = imod(a, b)} is equivalent to \code{a \%= b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{imul}{a, b}
\funcline{__imul__}{a, b}
\code{a = imul(a, b)} is equivalent to \code{a *= b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{ior}{a, b}
\funcline{__ior__}{a, b}
\code{a = ior(a, b)} is equivalent to \code{a |= b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{ipow}{a, b}
\funcline{__ipow__}{a, b}
\code{a = ipow(a, b)} is equivalent to \code{a **= b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{irepeat}{a, b}
\funcline{__irepeat__}{a, b}
\code{a = irepeat(a, b)} is equivalent to \code{a *= b} where
\var{a} is a sequence and \var{b} is an integer.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{irshift}{a, b}
\funcline{__irshift__}{a, b}
\code{a = irshift(a, b)} is equivalent to \code{a >>= b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{isub}{a, b}
\funcline{__isub__}{a, b}
\code{a = isub(a, b)} is equivalent to \code{a -= b}.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{itruediv}{a, b}
\funcline{__itruediv__}{a, b}
\code{a = itruediv(a, b)} is equivalent to \code{a /= b} when
\code{__future__.division} is in effect.
\versionadded{2.5}
\end{funcdesc}

\begin{funcdesc}{ixor}{a, b}
\funcline{__ixor__}{a, b}
\code{a = ixor(a, b)} is equivalent to \code{a \textasciicircum= b}.
\versionadded{2.5}
\end{funcdesc}


The \module{operator} module also defines a few predicates to test the
type of objects.  \note{Be careful not to misinterpret the
results of these functions; only \function{isCallable()} has any
measure of reliability with instance objects.  For example:}

\begin{verbatim}
>>> class C:
...     pass
... 
>>> import operator
>>> o = C()
>>> operator.isMappingType(o)
True
\end{verbatim}

\begin{funcdesc}{isCallable}{o}
\deprecated{2.0}{Use the \function{callable()} built-in function instead.}
Returns true if the object \var{o} can be called like a function,
otherwise it returns false.  True is returned for functions, bound and
unbound methods, class objects, and instance objects which support the
\method{__call__()} method.
\end{funcdesc}

\begin{funcdesc}{isMappingType}{o}
Returns true if the object \var{o} supports the mapping interface.
This is true for dictionaries and all instance objects defining
\method{__getitem__}.
\warning{There is no reliable way to test if an instance
supports the complete mapping protocol since the interface itself is
ill-defined.  This makes this test less useful than it otherwise might
be.}
\end{funcdesc}

\begin{funcdesc}{isNumberType}{o}
Returns true if the object \var{o} represents a number.  This is true
for all numeric types implemented in C.
\warning{There is no reliable way to test if an instance
supports the complete numeric interface since the interface itself is
ill-defined.  This makes this test less useful than it otherwise might
be.}
\end{funcdesc}

\begin{funcdesc}{isSequenceType}{o}
Returns true if the object \var{o} supports the sequence protocol.
This returns true for all objects which define sequence methods in C,
and for all instance objects defining \method{__getitem__}.
\warning{There is no reliable
way to test if an instance supports the complete sequence interface
since the interface itself is ill-defined.  This makes this test less
useful than it otherwise might be.}
\end{funcdesc}


Example: Build a dictionary that maps the ordinals from \code{0} to
\code{255} to their character equivalents.

\begin{verbatim}
>>> import operator
>>> d = {}
>>> keys = range(256)
>>> vals = map(chr, keys)
>>> map(operator.setitem, [d]*len(keys), keys, vals)
\end{verbatim}


The \module{operator} module also defines tools for generalized attribute
and item lookups.  These are useful for making fast field extractors
as arguments for \function{map()}, \function{sorted()},
\method{itertools.groupby()}, or other functions that expect a
function argument.

\begin{funcdesc}{attrgetter}{attr\optional{, args...}}
Return a callable object that fetches \var{attr} from its operand.
If more than one attribute is requested, returns a tuple of attributes.
After, \samp{f=attrgetter('name')}, the call \samp{f(b)} returns
\samp{b.name}.  After, \samp{f=attrgetter('name', 'date')}, the call
\samp{f(b)} returns \samp{(b.name, b.date)}. 
\versionadded{2.4}
\versionchanged[Added support for multiple attributes]{2.5}
\end{funcdesc}
    
\begin{funcdesc}{itemgetter}{item\optional{, args...}}
Return a callable object that fetches \var{item} from its operand.
If more than one item is requested, returns a tuple of items.
After, \samp{f=itemgetter(2)}, the call \samp{f(b)} returns
\samp{b[2]}.
After, \samp{f=itemgetter(2,5,3)}, the call \samp{f(b)} returns
\samp{(b[2], b[5], b[3])}.		
\versionadded{2.4}
\versionchanged[Added support for multiple item extraction]{2.5}		
\end{funcdesc}

Examples:
                
\begin{verbatim}
>>> from operator import itemgetter
>>> inventory = [('apple', 3), ('banana', 2), ('pear', 5), ('orange', 1)]
>>> getcount = itemgetter(1)
>>> map(getcount, inventory)
[3, 2, 5, 1]
>>> sorted(inventory, key=getcount)
[('orange', 1), ('banana', 2), ('apple', 3), ('pear', 5)]
\end{verbatim}
                

\subsection{Mapping Operators to Functions \label{operator-map}}

This table shows how abstract operations correspond to operator
symbols in the Python syntax and the functions in the
\refmodule{operator} module.


\begin{tableiii}{l|c|l}{textrm}{Operation}{Syntax}{Function}
  \lineiii{Addition}{\code{\var{a} + \var{b}}}
          {\code{add(\var{a}, \var{b})}}
  \lineiii{Concatenation}{\code{\var{seq1} + \var{seq2}}}
          {\code{concat(\var{seq1}, \var{seq2})}}
  \lineiii{Containment Test}{\code{\var{o} in \var{seq}}}
          {\code{contains(\var{seq}, \var{o})}}
  \lineiii{Division}{\code{\var{a} / \var{b}}}
          {\code{div(\var{a}, \var{b}) \#} without \code{__future__.division}}
  \lineiii{Division}{\code{\var{a} / \var{b}}}
          {\code{truediv(\var{a}, \var{b}) \#} with \code{__future__.division}}
  \lineiii{Division}{\code{\var{a} // \var{b}}}
          {\code{floordiv(\var{a}, \var{b})}}
  \lineiii{Bitwise And}{\code{\var{a} \&\ \var{b}}}
          {\code{and_(\var{a}, \var{b})}}
  \lineiii{Bitwise Exclusive Or}{\code{\var{a} \^\ \var{b}}}
          {\code{xor(\var{a}, \var{b})}}
  \lineiii{Bitwise Inversion}{\code{\~{} \var{a}}}
          {\code{invert(\var{a})}}
  \lineiii{Bitwise Or}{\code{\var{a} | \var{b}}}
          {\code{or_(\var{a}, \var{b})}}
  \lineiii{Exponentiation}{\code{\var{a} ** \var{b}}}
          {\code{pow(\var{a}, \var{b})}}
  \lineiii{Identity}{\code{\var{a} is \var{b}}}
          {\code{is_(\var{a}, \var{b})}}
  \lineiii{Identity}{\code{\var{a} is not \var{b}}}
          {\code{is_not(\var{a}, \var{b})}}
  \lineiii{Indexed Assignment}{\code{\var{o}[\var{k}] = \var{v}}}
          {\code{setitem(\var{o}, \var{k}, \var{v})}}
  \lineiii{Indexed Deletion}{\code{del \var{o}[\var{k}]}}
          {\code{delitem(\var{o}, \var{k})}}
  \lineiii{Indexing}{\code{\var{o}[\var{k}]}}
          {\code{getitem(\var{o}, \var{k})}}
  \lineiii{Left Shift}{\code{\var{a} <\code{<} \var{b}}}
          {\code{lshift(\var{a}, \var{b})}}
  \lineiii{Modulo}{\code{\var{a} \%\ \var{b}}}
          {\code{mod(\var{a}, \var{b})}}
  \lineiii{Multiplication}{\code{\var{a} * \var{b}}}
          {\code{mul(\var{a}, \var{b})}}
  \lineiii{Negation (Arithmetic)}{\code{- \var{a}}}
          {\code{neg(\var{a})}}
  \lineiii{Negation (Logical)}{\code{not \var{a}}}
          {\code{not_(\var{a})}}
  \lineiii{Right Shift}{\code{\var{a} >> \var{b}}}
          {\code{rshift(\var{a}, \var{b})}}
  \lineiii{Sequence Repitition}{\code{\var{seq} * \var{i}}}
          {\code{repeat(\var{seq}, \var{i})}}
  \lineiii{Slice Assignment}{\code{\var{seq}[\var{i}:\var{j}]} = \var{values}}
          {\code{setslice(\var{seq}, \var{i}, \var{j}, \var{values})}}
  \lineiii{Slice Deletion}{\code{del \var{seq}[\var{i}:\var{j}]}}
          {\code{delslice(\var{seq}, \var{i}, \var{j})}}
  \lineiii{Slicing}{\code{\var{seq}[\var{i}:\var{j}]}}
          {\code{getslice(\var{seq}, \var{i}, \var{j})}}
  \lineiii{String Formatting}{\code{\var{s} \%\ \var{o}}}
          {\code{mod(\var{s}, \var{o})}}
  \lineiii{Subtraction}{\code{\var{a} - \var{b}}}
          {\code{sub(\var{a}, \var{b})}}
  \lineiii{Truth Test}{\code{\var{o}}}
          {\code{truth(\var{o})}}
  \lineiii{Ordering}{\code{\var{a} < \var{b}}}
          {\code{lt(\var{a}, \var{b})}}
  \lineiii{Ordering}{\code{\var{a} <= \var{b}}}
          {\code{le(\var{a}, \var{b})}}
  \lineiii{Equality}{\code{\var{a} == \var{b}}}
          {\code{eq(\var{a}, \var{b})}}
  \lineiii{Difference}{\code{\var{a} != \var{b}}}
          {\code{ne(\var{a}, \var{b})}}
  \lineiii{Ordering}{\code{\var{a} >= \var{b}}}
          {\code{ge(\var{a}, \var{b})}}
  \lineiii{Ordering}{\code{\var{a} > \var{b}}}
          {\code{gt(\var{a}, \var{b})}}
\end{tableiii}