Doc/lib/libre.tex

\section{\module{re} ---
         Regular expression operations}
\declaremodule{standard}{re}
\moduleauthor{Fredrik Lundh}{fredrik@pythonware.com}
\sectionauthor{Andrew M. Kuchling}{amk@amk.ca}


\modulesynopsis{Regular expression search and match operations with a
                Perl-style expression syntax.}


This module provides regular expression matching operations similar to
those found in Perl.  Regular expression pattern strings may not
contain null bytes, but can specify the null byte using the
\code{\e\var{number}} notation.  Both patterns and strings to be
searched can be Unicode strings as well as 8-bit strings.  The
\module{re} module is always available.

Regular expressions use the backslash character (\character{\e}) to
indicate special forms or to allow special characters to be used
without invoking their special meaning.  This collides with Python's
usage of the same character for the same purpose in string literals;
for example, to match a literal backslash, one might have to write
\code{'\e\e\e\e'} as the pattern string, because the regular expression
must be \samp{\e\e}, and each backslash must be expressed as
\samp{\e\e} inside a regular Python string literal.

The solution is to use Python's raw string notation for regular
expression patterns; backslashes are not handled in any special way in
a string literal prefixed with \character{r}.  So \code{r"\e n"} is a
two-character string containing \character{\e} and \character{n},
while \code{"\e n"} is a one-character string containing a newline.
Usually patterns will be expressed in Python code using this raw
string notation.

\begin{seealso}
  \seetitle{Mastering Regular Expressions}{Book on regular expressions
            by Jeffrey Friedl, published by O'Reilly.  The second 
            edition of the book no longer covers Python at all, 
            but the first edition covered writing good regular expression
            patterns in great detail.}
\end{seealso}


\subsection{Regular Expression Syntax \label{re-syntax}}

A regular expression (or RE) specifies a set of strings that matches
it; the functions in this module let you check if a particular string
matches a given regular expression (or if a given regular expression
matches a particular string, which comes down to the same thing).

Regular expressions can be concatenated to form new regular
expressions; if \emph{A} and \emph{B} are both regular expressions,
then \emph{AB} is also a regular expression.  In general, if a string
\emph{p} matches \emph{A} and another string \emph{q} matches \emph{B},
the string \emph{pq} will match AB.  This holds unless \emph{A} or
\emph{B} contain low precedence operations; boundary conditions between
\emph{A} and \emph{B}; or have numbered group references.  Thus, complex
expressions can easily be constructed from simpler primitive
expressions like the ones described here.  For details of the theory
and implementation of regular expressions, consult the Friedl book
referenced above, or almost any textbook about compiler construction.

A brief explanation of the format of regular expressions follows.  For
further information and a gentler presentation, consult the Regular
Expression HOWTO, accessible from \url{http://www.python.org/doc/howto/}.

Regular expressions can contain both special and ordinary characters.
Most ordinary characters, like \character{A}, \character{a}, or
\character{0}, are the simplest regular expressions; they simply match
themselves.  You can concatenate ordinary characters, so \regexp{last}
matches the string \code{'last'}.  (In the rest of this section, we'll
write RE's in \regexp{this special style}, usually without quotes, and
strings to be matched \code{'in single quotes'}.)

Some characters, like \character{|} or \character{(}, are special.
Special characters either stand for classes of ordinary characters, or
affect how the regular expressions around them are interpreted.

The special characters are:
%
\begin{description}

\item[\character{.}] (Dot.)  In the default mode, this matches any
character except a newline.  If the \constant{DOTALL} flag has been
specified, this matches any character including a newline.

\item[\character{\textasciicircum}] (Caret.)  Matches the start of the
string, and in \constant{MULTILINE} mode also matches immediately
after each newline.

\item[\character{\$}] Matches the end of the string or just before the
newline at the end of the string, and in \constant{MULTILINE} mode
also matches before a newline.  \regexp{foo} matches both 'foo' and
'foobar', while the regular expression \regexp{foo\$} matches only
'foo'.  More interestingly, searching for \regexp{foo.\$} in
'foo1\textbackslash nfoo2\textbackslash n' matches 'foo2' normally,
but 'foo1' in \constant{MULTILINE} mode.

\item[\character{*}] Causes the resulting RE to
match 0 or more repetitions of the preceding RE, as many repetitions
as are possible.  \regexp{ab*} will
match 'a', 'ab', or 'a' followed by any number of 'b's.

\item[\character{+}] Causes the
resulting RE to match 1 or more repetitions of the preceding RE.
\regexp{ab+} will match 'a' followed by any non-zero number of 'b's; it
will not match just 'a'.

\item[\character{?}] Causes the resulting RE to
match 0 or 1 repetitions of the preceding RE.  \regexp{ab?} will
match either 'a' or 'ab'.

\item[\code{*?}, \code{+?}, \code{??}] The \character{*},
\character{+}, and \character{?} qualifiers are all \dfn{greedy}; they
match as much text as possible.  Sometimes this behaviour isn't
desired; if the RE \regexp{<.*>} is matched against
\code{'<H1>title</H1>'}, it will match the entire string, and not just
\code{'<H1>'}.  Adding \character{?} after the qualifier makes it
perform the match in \dfn{non-greedy} or \dfn{minimal} fashion; as
\emph{few} characters as possible will be matched.  Using \regexp{.*?}
in the previous expression will match only \code{'<H1>'}.

\item[\code{\{\var{m}\}}]
Specifies that exactly \var{m} copies of the previous RE should be
matched; fewer matches cause the entire RE not to match.  For example,
\regexp{a\{6\}} will match exactly six \character{a} characters, but
not five.

\item[\code{\{\var{m},\var{n}\}}] Causes the resulting RE to match from
\var{m} to \var{n} repetitions of the preceding RE, attempting to
match as many repetitions as possible.  For example, \regexp{a\{3,5\}}
will match from 3 to 5 \character{a} characters.  Omitting \var{m}
specifies a lower bound of zero, 
and omitting \var{n} specifies an infinite upper bound.  As an
example, \regexp{a\{4,\}b} will match \code{aaaab} or a thousand
\character{a} characters followed by a \code{b}, but not \code{aaab}.
The comma may not be omitted or the modifier would be confused with
the previously described form.

\item[\code{\{\var{m},\var{n}\}?}] Causes the resulting RE to
match from \var{m} to \var{n} repetitions of the preceding RE,
attempting to match as \emph{few} repetitions as possible.  This is
the non-greedy version of the previous qualifier.  For example, on the
6-character string \code{'aaaaaa'}, \regexp{a\{3,5\}} will match 5
\character{a} characters, while \regexp{a\{3,5\}?} will only match 3
characters.

\item[\character{\e}] Either escapes special characters (permitting
you to match characters like \character{*}, \character{?}, and so
forth), or signals a special sequence; special sequences are discussed
below.

If you're not using a raw string to
express the pattern, remember that Python also uses the
backslash as an escape sequence in string literals; if the escape
sequence isn't recognized by Python's parser, the backslash and
subsequent character are included in the resulting string.  However,
if Python would recognize the resulting sequence, the backslash should
be repeated twice.  This is complicated and hard to understand, so
it's highly recommended that you use raw strings for all but the
simplest expressions.

\item[\code{[]}] Used to indicate a set of characters.  Characters can
be listed individually, or a range of characters can be indicated by
giving two characters and separating them by a \character{-}.  Special
characters are not active inside sets.  For example, \regexp{[akm\$]}
will match any of the characters \character{a}, \character{k},
\character{m}, or \character{\$}; \regexp{[a-z]}
will match any lowercase letter, and \code{[a-zA-Z0-9]} matches any
letter or digit.  Character classes such as \code{\e w} or \code{\e S}
(defined below) are also acceptable inside a range.  If you want to
include a \character{]} or a \character{-} inside a set, precede it with a
backslash, or place it as the first character.  The
pattern \regexp{[]]} will match \code{']'}, for example.

You can match the characters not within a range by \dfn{complementing}
the set.  This is indicated by including a
\character{\textasciicircum} as the first character of the set;
\character{\textasciicircum} elsewhere will simply match the
\character{\textasciicircum} character.  For example,
\regexp{[{\textasciicircum}5]} will match
any character except \character{5}, and
\regexp{[\textasciicircum\code{\textasciicircum}]} will match any character
except \character{\textasciicircum}.

\item[\character{|}]\code{A|B}, where A and B can be arbitrary REs,
creates a regular expression that will match either A or B.  An
arbitrary number of REs can be separated by the \character{|} in this
way.  This can be used inside groups (see below) as well.  As the target
string is scanned, REs separated by \character{|} are tried from left to
right. When one pattern completely matches, that branch is accepted.
This means that once \code{A} matches, \code{B} will not be tested further,
even if it would produce a longer overall match.  In other words, the
\character{|} operator is never greedy.  To match a literal \character{|},
use \regexp{\e|}, or enclose it inside a character class, as in \regexp{[|]}.

\item[\code{(...)}] Matches whatever regular expression is inside the
parentheses, and indicates the start and end of a group; the contents
of a group can be retrieved after a match has been performed, and can
be matched later in the string with the \regexp{\e \var{number}} special
sequence, described below.  To match the literals \character{(} or
\character{)}, use \regexp{\e(} or \regexp{\e)}, or enclose them
inside a character class: \regexp{[(] [)]}.

\item[\code{(?...)}] This is an extension notation (a \character{?}
following a \character{(} is not meaningful otherwise).  The first
character after the \character{?}
determines what the meaning and further syntax of the construct is.
Extensions usually do not create a new group;
\regexp{(?P<\var{name}>...)} is the only exception to this rule.
Following are the currently supported extensions.

\item[\code{(?iLmsux)}] (One or more letters from the set \character{i},
\character{L}, \character{m}, \character{s}, \character{u},
\character{x}.)  The group matches the empty string; the letters set
the corresponding flags (\constant{re.I}, \constant{re.L},
\constant{re.M}, \constant{re.S}, \constant{re.U}, \constant{re.X})
for the entire regular expression.  This is useful if you wish to
include the flags as part of the regular expression, instead of
passing a \var{flag} argument to the \function{compile()} function.

Note that the \regexp{(?x)} flag changes how the expression is parsed.
It should be used first in the expression string, or after one or more
whitespace characters.  If there are non-whitespace characters before
the flag, the results are undefined.

\item[\code{(?:...)}] A non-grouping version of regular parentheses.
Matches whatever regular expression is inside the parentheses, but the
substring matched by the
group \emph{cannot} be retrieved after performing a match or
referenced later in the pattern.

\item[\code{(?P<\var{name}>...)}] Similar to regular parentheses, but
the substring matched by the group is accessible via the symbolic group
name \var{name}.  Group names must be valid Python identifiers, and
each group name must be defined only once within a regular expression.  A
symbolic group is also a numbered group, just as if the group were not
named.  So the group named 'id' in the example above can also be
referenced as the numbered group 1.

For example, if the pattern is
\regexp{(?P<id>[a-zA-Z_]\e w*)}, the group can be referenced by its
name in arguments to methods of match objects, such as
\code{m.group('id')} or \code{m.end('id')}, and also by name in
pattern text (for example, \regexp{(?P=id)}) and replacement text
(such as \code{\e g<id>}).

\item[\code{(?P=\var{name})}] Matches whatever text was matched by the
earlier group named \var{name}.

\item[\code{(?\#...)}] A comment; the contents of the parentheses are
simply ignored.

\item[\code{(?=...)}] Matches if \regexp{...} matches next, but doesn't
consume any of the string.  This is called a lookahead assertion.  For
example, \regexp{Isaac (?=Asimov)} will match \code{'Isaac~'} only if it's
followed by \code{'Asimov'}.

\item[\code{(?!...)}] Matches if \regexp{...} doesn't match next.  This
is a negative lookahead assertion.  For example,
\regexp{Isaac (?!Asimov)} will match \code{'Isaac~'} only if it's \emph{not}
followed by \code{'Asimov'}.

\item[\code{(?<=...)}] Matches if the current position in the string
is preceded by a match for \regexp{...} that ends at the current
position.  This is called a \dfn{positive lookbehind assertion}.
\regexp{(?<=abc)def} will find a match in \samp{abcdef}, since the
lookbehind will back up 3 characters and check if the contained
pattern matches.  The contained pattern must only match strings of
some fixed length, meaning that \regexp{abc} or \regexp{a|b} are
allowed, but \regexp{a*} and \regexp{a\{3,4\}} are not.  Note that
patterns which start with positive lookbehind assertions will never
match at the beginning of the string being searched; you will most
likely want to use the \function{search()} function rather than the
\function{match()} function:

\begin{verbatim}
>>> import re
>>> m = re.search('(?<=abc)def', 'abcdef')
>>> m.group(0)
'def'
\end{verbatim}

This example looks for a word following a hyphen:

\begin{verbatim}
>>> m = re.search('(?<=-)\w+', 'spam-egg')
>>> m.group(0)
'egg'
\end{verbatim}

\item[\code{(?<!...)}] Matches if the current position in the string
is not preceded by a match for \regexp{...}.  This is called a
\dfn{negative lookbehind assertion}.  Similar to positive lookbehind
assertions, the contained pattern must only match strings of some
fixed length.  Patterns which start with negative lookbehind
assertions may match at the beginning of the string being searched.

\item[\code{(?(\var{id/name})yes-pattern|no-pattern)}] Will try to match
with \regexp{yes-pattern} if the group with given \var{id} or \var{name}
exists, and with \regexp{no-pattern} if it doesn't. \regexp{|no-pattern}
is optional and can be omitted. For example, 
\regexp{(<)?(\e w+@\e w+(?:\e .\e w+)+)(?(1)>)} is a poor email matching
pattern, which will match with \code{'<user@host.com>'} as well as
\code{'user@host.com'}, but not with \code{'<user@host.com'}.
\versionadded{2.4}

\end{description}

The special sequences consist of \character{\e} and a character from the
list below.  If the ordinary character is not on the list, then the
resulting RE will match the second character.  For example,
\regexp{\e\$} matches the character \character{\$}.
%
\begin{description}

\item[\code{\e \var{number}}] Matches the contents of the group of the
same number.  Groups are numbered starting from 1.  For example,
\regexp{(.+) \e 1} matches \code{'the the'} or \code{'55 55'}, but not
\code{'the end'} (note
the space after the group).  This special sequence can only be used to
match one of the first 99 groups.  If the first digit of \var{number}
is 0, or \var{number} is 3 octal digits long, it will not be interpreted
as a group match, but as the character with octal value \var{number}.
Inside the \character{[} and \character{]} of a character class, all numeric
escapes are treated as characters.

\item[\code{\e A}] Matches only at the start of the string.

\item[\code{\e b}] Matches the empty string, but only at the
beginning or end of a word.  A word is defined as a sequence of
alphanumeric or underscore characters, so the end of a word is indicated by
whitespace or a non-alphanumeric, non-underscore character.  Note that 
{}\code{\e b} is defined as the boundary between \code{\e w} and \code{\e
W}, so the precise set of characters deemed to be alphanumeric depends on the
values of the \code{UNICODE} and \code{LOCALE} flags.  Inside a character
range, \regexp{\e b} represents the backspace character, for compatibility
with Python's string literals.

\item[\code{\e B}] Matches the empty string, but only when it is \emph{not}
at the beginning or end of a word.  This is just the opposite of {}\code{\e
b}, so is also subject to the settings of \code{LOCALE} and \code{UNICODE}.

\item[\code{\e d}]When the \constant{UNICODE} flag is not specified, matches
any decimal digit; this is equivalent to the set \regexp{[0-9]}. 
With \constant{UNICODE}, it will match whatever is classified as a digit
in the Unicode character properties database.

\item[\code{\e D}]When the \constant{UNICODE} flag is not specified, matches
any non-digit character; this is equivalent to the set 
\regexp{[{\textasciicircum}0-9]}.  With \constant{UNICODE}, it will match 
anything other than character marked as digits in the Unicode character 
properties database.

\item[\code{\e s}]When the \constant{LOCALE} and \constant{UNICODE}
flags are not specified, matches any whitespace character; this is
equivalent to the set \regexp{[ \e t\e n\e r\e f\e v]}.
With \constant{LOCALE}, it will match this set plus whatever characters
are defined as space for the current locale. If \constant{UNICODE} is set,
this will match the characters \regexp{[ \e t\e n\e r\e f\e v]} plus
whatever is classified as space in the Unicode character properties
database.

\item[\code{\e S}]When the \constant{LOCALE} and \constant{UNICODE}
flags are not specified, matches any non-whitespace character; this is
equivalent to the set \regexp{[\textasciicircum\ \e t\e n\e r\e f\e v]}
With \constant{LOCALE}, it will match any character not in this set,
and not defined as space in the current locale. If \constant{UNICODE}
is set, this will match anything other than \regexp{[ \e t\e n\e r\e f\e v]}
and characters marked as space in the Unicode character properties database.

\item[\code{\e w}]When the \constant{LOCALE} and \constant{UNICODE}
flags are not specified, matches any alphanumeric character and the
underscore; this is equivalent to the set
\regexp{[a-zA-Z0-9_]}.  With \constant{LOCALE}, it will match the set
\regexp{[0-9_]} plus whatever characters are defined as alphanumeric for
the current locale.  If \constant{UNICODE} is set, this will match the
characters \regexp{[0-9_]} plus whatever is classified as alphanumeric
in the Unicode character properties database.

\item[\code{\e W}]When the \constant{LOCALE} and \constant{UNICODE}
flags are not specified, matches any non-alphanumeric character; this
is equivalent to the set \regexp{[{\textasciicircum}a-zA-Z0-9_]}.   With
\constant{LOCALE}, it will match any character not in the set
\regexp{[0-9_]}, and not defined as alphanumeric for the current locale.
If \constant{UNICODE} is set, this will match anything other than
\regexp{[0-9_]} and characters marked as alphanumeric in the Unicode
character properties database.

\item[\code{\e Z}]Matches only at the end of the string.

\end{description}

Most of the standard escapes supported by Python string literals are
also accepted by the regular expression parser:

\begin{verbatim}
\a      \b      \f      \n
\r      \t      \v      \x
\\
\end{verbatim}

Octal escapes are included in a limited form: If the first digit is a
0, or if there are three octal digits, it is considered an octal
escape. Otherwise, it is a group reference.  As for string literals,
octal escapes are always at most three digits in length.


% Note the lack of a period in the section title; it causes problems
% with readers of the GNU info version.  See http://www.python.org/sf/581414.
\subsection{Matching vs Searching \label{matching-searching}}
\sectionauthor{Fred L. Drake, Jr.}{fdrake@acm.org}

Python offers two different primitive operations based on regular
expressions: match and search.  If you are accustomed to Perl's
semantics, the search operation is what you're looking for.  See the
\function{search()} function and corresponding method of compiled
regular expression objects.

Note that match may differ from search using a regular expression
beginning with \character{\textasciicircum}:
\character{\textasciicircum} matches only at the
start of the string, or in \constant{MULTILINE} mode also immediately
following a newline.  The ``match'' operation succeeds only if the
pattern matches at the start of the string regardless of mode, or at
the starting position given by the optional \var{pos} argument
regardless of whether a newline precedes it.

% Examples from Tim Peters:
\begin{verbatim}
re.compile("a").match("ba", 1)           # succeeds
re.compile("^a").search("ba", 1)         # fails; 'a' not at start
re.compile("^a").search("\na", 1)        # fails; 'a' not at start
re.compile("^a", re.M).search("\na", 1)  # succeeds
re.compile("^a", re.M).search("ba", 1)   # fails; no preceding \n
\end{verbatim}


\subsection{Module Contents}
\nodename{Contents of Module re}

The module defines several functions, constants, and an exception. Some of the
functions are simplified versions of the full featured methods for compiled
regular expressions.  Most non-trivial applications always use the compiled
form.

\begin{funcdesc}{compile}{pattern\optional{, flags}}
  Compile a regular expression pattern into a regular expression
  object, which can be used for matching using its \function{match()} and
  \function{search()} methods, described below.

  The expression's behaviour can be modified by specifying a
  \var{flags} value.  Values can be any of the following variables,
  combined using bitwise OR (the \code{|} operator).

The sequence

\begin{verbatim}
prog = re.compile(pat)
result = prog.match(str)
\end{verbatim}

is equivalent to

\begin{verbatim}
result = re.match(pat, str)
\end{verbatim}

but the version using \function{compile()} is more efficient when the
expression will be used several times in a single program.
%(The compiled version of the last pattern passed to
%\function{re.match()} or \function{re.search()} is cached, so
%programs that use only a single regular expression at a time needn't
%worry about compiling regular expressions.)
\end{funcdesc}

\begin{datadesc}{I}
\dataline{IGNORECASE}
Perform case-insensitive matching; expressions like \regexp{[A-Z]}
will match lowercase letters, too.  This is not affected by the
current locale.
\end{datadesc}

\begin{datadesc}{L}
\dataline{LOCALE}
Make \regexp{\e w}, \regexp{\e W}, \regexp{\e b}, \regexp{\e B},
\regexp{\e s} and \regexp{\e S} dependent on the current locale.
\end{datadesc}

\begin{datadesc}{M}
\dataline{MULTILINE}
When specified, the pattern character \character{\textasciicircum}
matches at the beginning of the string and at the beginning of each
line (immediately following each newline); and the pattern character
\character{\$} matches at the end of the string and at the end of each
line (immediately preceding each newline).  By default,
\character{\textasciicircum} matches only at the beginning of the
string, and \character{\$} only at the end of the string and
immediately before the newline (if any) at the end of the string.
\end{datadesc}

\begin{datadesc}{S}
\dataline{DOTALL}
Make the \character{.} special character match any character at all,
including a newline; without this flag, \character{.} will match
anything \emph{except} a newline.
\end{datadesc}

\begin{datadesc}{U}
\dataline{UNICODE}
Make \regexp{\e w}, \regexp{\e W}, \regexp{\e b}, \regexp{\e B},
\regexp{\e d}, \regexp{\e D}, \regexp{\e s} and \regexp{\e S}
dependent on the Unicode character properties database.
\versionadded{2.0}
\end{datadesc}

\begin{datadesc}{X}
\dataline{VERBOSE}
This flag allows you to write regular expressions that look nicer.
Whitespace within the pattern is ignored,
except when in a character class or preceded by an unescaped
backslash, and, when a line contains a \character{\#} neither in a
character class or preceded by an unescaped backslash, all characters
from the leftmost such \character{\#} through the end of the line are
ignored.
% XXX should add an example here
\end{datadesc}


\begin{funcdesc}{search}{pattern, string\optional{, flags}}
  Scan through \var{string} looking for a location where the regular
  expression \var{pattern} produces a match, and return a
  corresponding \class{MatchObject} instance.
  Return \code{None} if no
  position in the string matches the pattern; note that this is
  different from finding a zero-length match at some point in the string.
\end{funcdesc}

\begin{funcdesc}{match}{pattern, string\optional{, flags}}
  If zero or more characters at the beginning of \var{string} match
  the regular expression \var{pattern}, return a corresponding
  \class{MatchObject} instance.  Return \code{None} if the string does not
  match the pattern; note that this is different from a zero-length
  match.

  \note{If you want to locate a match anywhere in
  \var{string}, use \method{search()} instead.}
\end{funcdesc}

\begin{funcdesc}{split}{pattern, string\optional{, maxsplit\code{ = 0}}}
  Split \var{string} by the occurrences of \var{pattern}.  If
  capturing parentheses are used in \var{pattern}, then the text of all
  groups in the pattern are also returned as part of the resulting list.
  If \var{maxsplit} is nonzero, at most \var{maxsplit} splits
  occur, and the remainder of the string is returned as the final
  element of the list.  (Incompatibility note: in the original Python
  1.5 release, \var{maxsplit} was ignored.  This has been fixed in
  later releases.)

\begin{verbatim}
>>> re.split('\W+', 'Words, words, words.')
['Words', 'words', 'words', '']
>>> re.split('(\W+)', 'Words, words, words.')
['Words', ', ', 'words', ', ', 'words', '.', '']
>>> re.split('\W+', 'Words, words, words.', 1)
['Words', 'words, words.']
\end{verbatim}

  This function combines and extends the functionality of
  the old \function{regsub.split()} and \function{regsub.splitx()}.
\end{funcdesc}

\begin{funcdesc}{findall}{pattern, string\optional{, flags}}
  Return a list of all non-overlapping matches of \var{pattern} in
  \var{string}.  If one or more groups are present in the pattern,
  return a list of groups; this will be a list of tuples if the
  pattern has more than one group.  Empty matches are included in the
  result unless they touch the beginning of another match.
  \versionadded{1.5.2}
  \versionchanged[Added the optional flags argument]{2.4}
\end{funcdesc}

\begin{funcdesc}{finditer}{pattern, string\optional{, flags}}
  Return an iterator over all non-overlapping matches for the RE
  \var{pattern} in \var{string}.  For each match, the iterator returns
  a match object.  Empty matches are included in the result unless they
  touch the beginning of another match.
  \versionadded{2.2}
  \versionchanged[Added the optional flags argument]{2.4}
\end{funcdesc}

\begin{funcdesc}{sub}{pattern, repl, string\optional{, count}}
  Return the string obtained by replacing the leftmost non-overlapping
  occurrences of \var{pattern} in \var{string} by the replacement
  \var{repl}.  If the pattern isn't found, \var{string} is returned
  unchanged.  \var{repl} can be a string or a function; if it is a
  string, any backslash escapes in it are processed.  That is,
  \samp{\e n} is converted to a single newline character, \samp{\e r}
  is converted to a linefeed, and so forth.  Unknown escapes such as
  \samp{\e j} are left alone.  Backreferences, such as \samp{\e6}, are
  replaced with the substring matched by group 6 in the pattern.  For
  example:

\begin{verbatim}
>>> re.sub(r'def\s+([a-zA-Z_][a-zA-Z_0-9]*)\s*\(\s*\):',
...        r'static PyObject*\npy_\1(void)\n{',
...        'def myfunc():')
'static PyObject*\npy_myfunc(void)\n{'
\end{verbatim}

  If \var{repl} is a function, it is called for every non-overlapping
  occurrence of \var{pattern}.  The function takes a single match
  object argument, and returns the replacement string.  For example:

\begin{verbatim}
>>> def dashrepl(matchobj):
...     if matchobj.group(0) == '-': return ' '
...     else: return '-'
>>> re.sub('-{1,2}', dashrepl, 'pro----gram-files')
'pro--gram files'
\end{verbatim}

  The pattern may be a string or an RE object; if you need to specify
  regular expression flags, you must use a RE object, or use embedded
  modifiers in a pattern; for example, \samp{sub("(?i)b+", "x", "bbbb
  BBBB")} returns \code{'x x'}.

  The optional argument \var{count} is the maximum number of pattern
  occurrences to be replaced; \var{count} must be a non-negative
  integer.  If omitted or zero, all occurrences will be replaced.
  Empty matches for the pattern are replaced only when not adjacent to
  a previous match, so \samp{sub('x*', '-', 'abc')} returns
  \code{'-a-b-c-'}.

  In addition to character escapes and backreferences as described
  above, \samp{\e g<name>} will use the substring matched by the group
  named \samp{name}, as defined by the \regexp{(?P<name>...)} syntax.
  \samp{\e g<number>} uses the corresponding group number;
  \samp{\e g<2>} is therefore equivalent to \samp{\e 2}, but isn't
  ambiguous in a replacement such as \samp{\e g<2>0}.  \samp{\e 20}
  would be interpreted as a reference to group 20, not a reference to
  group 2 followed by the literal character \character{0}.  The
  backreference \samp{\e g<0>} substitutes in the entire substring
  matched by the RE.
\end{funcdesc}

\begin{funcdesc}{subn}{pattern, repl, string\optional{, count}}
  Perform the same operation as \function{sub()}, but return a tuple
  \code{(\var{new_string}, \var{number_of_subs_made})}.
\end{funcdesc}

\begin{funcdesc}{escape}{string}
  Return \var{string} with all non-alphanumerics backslashed; this is
  useful if you want to match an arbitrary literal string that may have
  regular expression metacharacters in it.
\end{funcdesc}

\begin{excdesc}{error}
  Exception raised when a string passed to one of the functions here
  is not a valid regular expression (for example, it might contain
  unmatched parentheses) or when some other error occurs during
  compilation or matching.  It is never an error if a string contains
  no match for a pattern.
\end{excdesc}


\subsection{Regular Expression Objects \label{re-objects}}

Compiled regular expression objects support the following methods and
attributes:

\begin{methoddesc}[RegexObject]{match}{string\optional{, pos\optional{,
                                       endpos}}}
  If zero or more characters at the beginning of \var{string} match
  this regular expression, return a corresponding
  \class{MatchObject} instance.  Return \code{None} if the string does not
  match the pattern; note that this is different from a zero-length
  match.

  \note{If you want to locate a match anywhere in
  \var{string}, use \method{search()} instead.}

  The optional second parameter \var{pos} gives an index in the string
  where the search is to start; it defaults to \code{0}.  This is not
  completely equivalent to slicing the string; the
  \code{'\textasciicircum'} pattern
  character matches at the real beginning of the string and at positions
  just after a newline, but not necessarily at the index where the search
  is to start.

  The optional parameter \var{endpos} limits how far the string will
  be searched; it will be as if the string is \var{endpos} characters
  long, so only the characters from \var{pos} to \code{\var{endpos} -
  1} will be searched for a match.  If \var{endpos} is less than
  \var{pos}, no match will be found, otherwise, if \var{rx} is a
  compiled regular expression object,
  \code{\var{rx}.match(\var{string}, 0, 50)} is equivalent to
  \code{\var{rx}.match(\var{string}[:50], 0)}.
\end{methoddesc}

\begin{methoddesc}[RegexObject]{search}{string\optional{, pos\optional{,
                                        endpos}}}
  Scan through \var{string} looking for a location where this regular
  expression produces a match, and return a
  corresponding \class{MatchObject} instance.  Return \code{None} if no
  position in the string matches the pattern; note that this is
  different from finding a zero-length match at some point in the string.

  The optional \var{pos} and \var{endpos} parameters have the same
  meaning as for the \method{match()} method.
\end{methoddesc}

\begin{methoddesc}[RegexObject]{split}{string\optional{,
                                       maxsplit\code{ = 0}}}
Identical to the \function{split()} function, using the compiled pattern.
\end{methoddesc}

\begin{methoddesc}[RegexObject]{findall}{string\optional{, pos\optional{,
                                        endpos}}}
Identical to the \function{findall()} function, using the compiled pattern.
\end{methoddesc}

\begin{methoddesc}[RegexObject]{finditer}{string\optional{, pos\optional{,
                                        endpos}}}
Identical to the \function{finditer()} function, using the compiled pattern.
\end{methoddesc}

\begin{methoddesc}[RegexObject]{sub}{repl, string\optional{, count\code{ = 0}}}
Identical to the \function{sub()} function, using the compiled pattern.
\end{methoddesc}

\begin{methoddesc}[RegexObject]{subn}{repl, string\optional{,
                                      count\code{ = 0}}}
Identical to the \function{subn()} function, using the compiled pattern.
\end{methoddesc}


\begin{memberdesc}[RegexObject]{flags}
The flags argument used when the RE object was compiled, or
\code{0} if no flags were provided.
\end{memberdesc}

\begin{memberdesc}[RegexObject]{groupindex}
A dictionary mapping any symbolic group names defined by
\regexp{(?P<\var{id}>)} to group numbers.  The dictionary is empty if no
symbolic groups were used in the pattern.
\end{memberdesc}

\begin{memberdesc}[RegexObject]{pattern}
The pattern string from which the RE object was compiled.
\end{memberdesc}


\subsection{Match Objects \label{match-objects}}

\class{MatchObject} instances support the following methods and
attributes:

\begin{methoddesc}[MatchObject]{expand}{template}
 Return the string obtained by doing backslash substitution on the
template string \var{template}, as done by the \method{sub()} method.
Escapes such as \samp{\e n} are converted to the appropriate
characters, and numeric backreferences (\samp{\e 1}, \samp{\e 2}) and
named backreferences (\samp{\e g<1>}, \samp{\e g<name>}) are replaced
by the contents of the corresponding group.
\end{methoddesc}

\begin{methoddesc}[MatchObject]{group}{\optional{group1, \moreargs}}
Returns one or more subgroups of the match.  If there is a single
argument, the result is a single string; if there are
multiple arguments, the result is a tuple with one item per argument.
Without arguments, \var{group1} defaults to zero (the whole match
is returned).
If a \var{groupN} argument is zero, the corresponding return value is the
entire matching string; if it is in the inclusive range [1..99], it is
the string matching the corresponding parenthesized group.  If a
group number is negative or larger than the number of groups defined
in the pattern, an \exception{IndexError} exception is raised.
If a group is contained in a part of the pattern that did not match,
the corresponding result is \code{None}.  If a group is contained in a
part of the pattern that matched multiple times, the last match is
returned.

If the regular expression uses the \regexp{(?P<\var{name}>...)} syntax,
the \var{groupN} arguments may also be strings identifying groups by
their group name.  If a string argument is not used as a group name in
the pattern, an \exception{IndexError} exception is raised.

A moderately complicated example:

\begin{verbatim}
m = re.match(r"(?P<int>\d+)\.(\d*)", '3.14')
\end{verbatim}

After performing this match, \code{m.group(1)} is \code{'3'}, as is
\code{m.group('int')}, and \code{m.group(2)} is \code{'14'}.
\end{methoddesc}

\begin{methoddesc}[MatchObject]{groups}{\optional{default}}
Return a tuple containing all the subgroups of the match, from 1 up to
however many groups are in the pattern.  The \var{default} argument is
used for groups that did not participate in the match; it defaults to
\code{None}.  (Incompatibility note: in the original Python 1.5
release, if the tuple was one element long, a string would be returned
instead.  In later versions (from 1.5.1 on), a singleton tuple is
returned in such cases.)
\end{methoddesc}

\begin{methoddesc}[MatchObject]{groupdict}{\optional{default}}
Return a dictionary containing all the \emph{named} subgroups of the
match, keyed by the subgroup name.  The \var{default} argument is
used for groups that did not participate in the match; it defaults to
\code{None}.
\end{methoddesc}

\begin{methoddesc}[MatchObject]{start}{\optional{group}}
\methodline{end}{\optional{group}}
Return the indices of the start and end of the substring
matched by \var{group}; \var{group} defaults to zero (meaning the whole
matched substring).
Return \code{-1} if \var{group} exists but
did not contribute to the match.  For a match object
\var{m}, and a group \var{g} that did contribute to the match, the
substring matched by group \var{g} (equivalent to
\code{\var{m}.group(\var{g})}) is

\begin{verbatim}
m.string[m.start(g):m.end(g)]
\end{verbatim}

Note that
\code{m.start(\var{group})} will equal \code{m.end(\var{group})} if
\var{group} matched a null string.  For example, after \code{\var{m} =
re.search('b(c?)', 'cba')}, \code{\var{m}.start(0)} is 1,
\code{\var{m}.end(0)} is 2, \code{\var{m}.start(1)} and
\code{\var{m}.end(1)} are both 2, and \code{\var{m}.start(2)} raises
an \exception{IndexError} exception.
\end{methoddesc}

\begin{methoddesc}[MatchObject]{span}{\optional{group}}
For \class{MatchObject} \var{m}, return the 2-tuple
\code{(\var{m}.start(\var{group}), \var{m}.end(\var{group}))}.
Note that if \var{group} did not contribute to the match, this is
\code{(-1, -1)}.  Again, \var{group} defaults to zero.
\end{methoddesc}

\begin{memberdesc}[MatchObject]{pos}
The value of \var{pos} which was passed to the \function{search()} or
\function{match()} method of the \class{RegexObject}.  This is the
index into the string at which the RE engine started looking for a
match.
\end{memberdesc}

\begin{memberdesc}[MatchObject]{endpos}
The value of \var{endpos} which was passed to the \function{search()}
or \function{match()} method of the \class{RegexObject}.  This is the
index into the string beyond which the RE engine will not go.
\end{memberdesc}

\begin{memberdesc}[MatchObject]{lastindex}
The integer index of the last matched capturing group, or \code{None}
if no group was matched at all. For example, the expressions
\regexp{(a)b}, \regexp{((a)(b))}, and \regexp{((ab))} will have
\code{lastindex == 1} if applied to the string \code{'ab'},
while the expression \regexp{(a)(b)} will have \code{lastindex == 2},
if applied to the same string.
\end{memberdesc}

\begin{memberdesc}[MatchObject]{lastgroup}
The name of the last matched capturing group, or \code{None} if the
group didn't have a name, or if no group was matched at all.
\end{memberdesc}

\begin{memberdesc}[MatchObject]{re}
The regular expression object whose \method{match()} or
\method{search()} method produced this \class{MatchObject} instance.
\end{memberdesc}

\begin{memberdesc}[MatchObject]{string}
The string passed to \function{match()} or \function{search()}.
\end{memberdesc}

\subsection{Examples}

\leftline{\strong{Simulating \cfunction{scanf()}}}

Python does not currently have an equivalent to \cfunction{scanf()}.
\ttindex{scanf()}
Regular expressions are generally more powerful, though also more
verbose, than \cfunction{scanf()} format strings.  The table below
offers some more-or-less equivalent mappings between
\cfunction{scanf()} format tokens and regular expressions.

\begin{tableii}{l|l}{textrm}{\cfunction{scanf()} Token}{Regular Expression}
  \lineii{\code{\%c}}
         {\regexp{.}}
  \lineii{\code{\%5c}}
         {\regexp{.\{5\}}}
  \lineii{\code{\%d}}
         {\regexp{[-+]?\e d+}}
  \lineii{\code{\%e}, \code{\%E}, \code{\%f}, \code{\%g}}
         {\regexp{[-+]?(\e d+(\e.\e d*)?|\e d*\e.\e d+)([eE][-+]?\e d+)?}}
  \lineii{\code{\%i}}
         {\regexp{[-+]?(0[xX][\e dA-Fa-f]+|0[0-7]*|\e d+)}}
  \lineii{\code{\%o}}
         {\regexp{0[0-7]*}}
  \lineii{\code{\%s}}
         {\regexp{\e S+}}
  \lineii{\code{\%u}}
         {\regexp{\e d+}}
  \lineii{\code{\%x}, \code{\%X}}
         {\regexp{0[xX][\e dA-Fa-f]+}}
\end{tableii}

To extract the filename and numbers from a string like

\begin{verbatim}
    /usr/sbin/sendmail - 0 errors, 4 warnings
\end{verbatim}

you would use a \cfunction{scanf()} format like

\begin{verbatim}
    %s - %d errors, %d warnings
\end{verbatim}

The equivalent regular expression would be

\begin{verbatim}
    (\S+) - (\d+) errors, (\d+) warnings
\end{verbatim}

\leftline{\strong{Avoiding recursion}}

If you create regular expressions that require the engine to perform a
lot of recursion, you may encounter a RuntimeError exception with
the message \code{maximum recursion limit} exceeded. For example,

\begin{verbatim}
>>> import re
>>> s = 'Begin ' + 1000*'a very long string ' + 'end'
>>> re.match('Begin (\w| )*? end', s).end()
Traceback (most recent call last):
  File "<stdin>", line 1, in ?
  File "/usr/local/lib/python2.3/sre.py", line 132, in match
    return _compile(pattern, flags).match(string)
RuntimeError: maximum recursion limit exceeded
\end{verbatim}

You can often restructure your regular expression to avoid recursion.

Starting with Python 2.3, simple uses of the \regexp{*?} pattern are
special-cased to avoid recursion.  Thus, the above regular expression
can avoid recursion by being recast as
\regexp{Begin [a-zA-Z0-9_ ]*?end}.  As a further benefit, such regular
expressions will run faster than their recursive equivalents.