path: root/pod/perlreref.pod

=head1 NAME

perlreref - Perl Regular Expressions Reference

=head1 DESCRIPTION

This is a quick reference to Perl's regular expressions.
For full information see L<perlre> and L<perlop>, as well
as the L</"SEE ALSO"> section in this document.

=head2 OPERATORS

C<=~> determines to which variable the regex is applied.
In its absence, $_ is used.

    $var =~ /foo/;

C<!~> determines to which variable the regex is applied,
and negates the result of the match; it returns
false if the match succeeds, and true if it fails.

    $var !~ /foo/;

C<m/pattern/msixpogc> searches a string for a pattern match,
applying the given options.

    m  Multiline mode - ^ and $ match internal lines
    s  match as a Single line - . matches \n
    i  case-Insensitive
    x  eXtended legibility - free whitespace and comments
    p  Preserve a copy of the matched string -
       ${^PREMATCH}, ${^MATCH}, ${^POSTMATCH} will be defined.
    o  compile pattern Once
    g  Global - all occurrences
    c  don't reset pos on failed matches when using /g

If 'pattern' is an empty string, the last I<successfully> matched
regex is used. Delimiters other than '/' may be used for both this
operator and the following ones. The leading C<m> can be omitted
if the delimiter is '/'.

C<qr/pattern/msixpo> lets you store a regex in a variable,
or pass one around. Modifiers as for C<m//>, and are stored
within the regex.

C<s/pattern/replacement/msixpogce> substitutes matches of
'pattern' with 'replacement'. Modifiers as for C<m//>,
with two additions:

    e  Evaluate 'replacement' as an expression
    r  Return substitution and leave the original string untouched.

'e' may be specified multiple times. 'replacement' is interpreted
as a double quoted string unless a single-quote (C<'>) is the delimiter.

C<?pattern?> is like C<m/pattern/> but matches only once. No alternate
delimiters can be used.  Must be reset with reset().

=head2 SYNTAX

 \       Escapes the character immediately following it
 .       Matches any single character except a newline (unless /s is
           used)
 ^       Matches at the beginning of the string (or line, if /m is used)
 $       Matches at the end of the string (or line, if /m is used)
 *       Matches the preceding element 0 or more times
 +       Matches the preceding element 1 or more times
 ?       Matches the preceding element 0 or 1 times
 {...}   Specifies a range of occurrences for the element preceding it
 [...]   Matches any one of the characters contained within the brackets
 (...)   Groups subexpressions for capturing to $1, $2...
 (?:...) Groups subexpressions without capturing (cluster)
 |       Matches either the subexpression preceding or following it
 \g1 or \g{1}, \g2 ...    Matches the text from the Nth group
 \1, \2, \3 ...           Matches the text from the Nth group
 \g-1 or \g{-1}, \g-2 ... Matches the text from the Nth previous group
 \g{name}     Named backreference
 \k<name>     Named backreference
 \k'name'     Named backreference
 (?P=name)    Named backreference (python syntax)

=head2 ESCAPE SEQUENCES

These work as in normal strings.

   \a       Alarm (beep)
   \e       Escape
   \f       Formfeed
   \n       Newline
   \r       Carriage return
   \t       Tab
   \037     Char whose ordinal is the 3 octal digits, max \777
   \o{2307} Char whose ordinal is the octal number, unrestricted
   \x7f     Char whose ordinal is the 2 hex digits, max \xFF
   \x{263a} Char whose ordinal is the hex number, unrestricted
   \cx      Control-x
   \N{name} A named Unicode character or character sequence
   \N{U+263D} A Unicode character by hex ordinal

   \l  Lowercase next character
   \u  Titlecase next character
   \L  Lowercase until \E
   \U  Uppercase until \E
   \Q  Disable pattern metacharacters until \E
   \E  End modification

For Titlecase, see L</Titlecase>.

This one works differently from normal strings:

   \b  An assertion, not backspace, except in a character class

=head2 CHARACTER CLASSES

   [amy]    Match 'a', 'm' or 'y'
   [f-j]    Dash specifies "range"
   [f-j-]   Dash escaped or at start or end means 'dash'
   [^f-j]   Caret indicates "match any character _except_ these"

The following sequences (except C<\N>) work within or without a character class.
The first six are locale aware, all are Unicode aware. See L<perllocale>
and L<perlunicode> for details.

   \d      A digit
   \D      A nondigit
   \w      A word character
   \W      A non-word character
   \s      A whitespace character
   \S      A non-whitespace character
   \h      An horizontal whitespace
   \H      A non horizontal whitespace
   \N      A non newline (when not followed by '{NAME}'; experimental;
           not valid in a character class; equivalent to [^\n]; it's
           like '.' without /s modifier)
   \v      A vertical whitespace
   \V      A non vertical whitespace
   \R      A generic newline           (?>\v|\x0D\x0A)

   \C      Match a byte (with Unicode, '.' matches a character)
   \pP     Match P-named (Unicode) property
   \p{...} Match Unicode property with name longer than 1 character
   \PP     Match non-P
   \P{...} Match lack of Unicode property with name longer than 1 char
   \X      Match Unicode extended grapheme cluster

POSIX character classes and their Unicode and Perl equivalents:

           ASCII-         Full-
           range          range   backslash
 POSIX    \p{...}         \p{}    sequence       Description
 -----------------------------------------------------------------------
 alnum   PosixAlnum       Alnum               Alpha plus Digit
 alpha   PosixAlpha       Alpha               Alphabetic characters
 ascii   ASCII                                Any ASCII character
 blank   PosixBlank       Blank     \h        Horizontal whitespace;
                                                full-range also written
                                                as \p{HorizSpace} (GNU
                                                extension)
 cntrl   PosixCntrl       Cntrl               Control characters
 digit   PosixDigit       Digit     \d        Decimal digits
 graph   PosixGraph       Graph               Alnum plus Punct
 lower   PosixLower       Lower               Lowercase characters
 print   PosixPrint       Print               Graph plus Print, but not
                                                any Cntrls
 punct   PosixPunct       Punct               These aren't precisely
                                                equivalent.  See NOTE,
                                                below.
 space   PosixSpace       Space     [\s\cK]   Whitespace
         PerlSpace        SpacePerl \s        Perl's whitespace
                                                definition
 upper   PosixUpper       Upper               Uppercase characters
 word    PerlWord         Word      \w        Alnum plus '_' (Perl
                                                extension)
 xdigit  ASCII_Hex_Digit  XDigit              Hexadecimal digit,
                                                ASCII-range is
                                                [0-9A-Fa-f]

NOTE on C<[[:punct:]]>, C<\p{PosixPunct}> and C<\p{Punct}>:
In the ASCII range, C<[[:punct:]]> and C<\p{PosixPunct}> match
C<[-!"#$%&'()*+,./:;<=E<gt>?@[\\\]^_`{|}~]> (although if a locale is in
effect, it could alter the behavior of C<[[:punct:]]>); and C<\p{Punct}>
matches C<[-!"#%&'()*,./:;?@[\\\]_{}]>.  When matching a UTF-8 string,
C<[[:punct:]]> matches what it does in the ASCII range, plus what
C<\p{Punct}> matches.  C<\p{Punct}> matches, anything that isn't a
control, an alphanumeric, a space, nor a symbol.

Within a character class:

    POSIX      traditional   Unicode
  [:digit:]       \d        \p{Digit}
  [:^digit:]      \D        \P{Digit}

=head2 ANCHORS

All are zero-width assertions.

   ^  Match string start (or line, if /m is used)
   $  Match string end (or line, if /m is used) or before newline
   \b Match word boundary (between \w and \W)
   \B Match except at word boundary (between \w and \w or \W and \W)
   \A Match string start (regardless of /m)
   \Z Match string end (before optional newline)
   \z Match absolute string end
   \G Match where previous m//g left off
   \K Keep the stuff left of the \K, don't include it in $&

=head2 QUANTIFIERS

Quantifiers are greedy by default and match the B<longest> leftmost.

   Maximal Minimal Possessive Allowed range
   ------- ------- ---------- -------------
   {n,m}   {n,m}?  {n,m}+     Must occur at least n times
                              but no more than m times
   {n,}    {n,}?   {n,}+      Must occur at least n times
   {n}     {n}?    {n}+       Must occur exactly n times
   *       *?      *+         0 or more times (same as {0,})
   +       +?      ++         1 or more times (same as {1,})
   ?       ??      ?+         0 or 1 time (same as {0,1})

The possessive forms (new in Perl 5.10) prevent backtracking: what gets
matched by a pattern with a possessive quantifier will not be backtracked
into, even if that causes the whole match to fail.

There is no quantifier C<{,n}>. That's interpreted as a literal string.

=head2 EXTENDED CONSTRUCTS

   (?#text)          A comment
   (?:...)           Groups subexpressions without capturing (cluster)
   (?pimsx-imsx:...) Enable/disable option (as per m// modifiers)
   (?=...)           Zero-width positive lookahead assertion
   (?!...)           Zero-width negative lookahead assertion
   (?<=...)          Zero-width positive lookbehind assertion
   (?<!...)          Zero-width negative lookbehind assertion
   (?>...)           Grab what we can, prohibit backtracking
   (?|...)           Branch reset
   (?<name>...)      Named capture
   (?'name'...)      Named capture
   (?P<name>...)     Named capture (python syntax)
   (?{ code })       Embedded code, return value becomes $^R
   (??{ code })      Dynamic regex, return value used as regex
   (?N)              Recurse into subpattern number N
   (?-N), (?+N)      Recurse into Nth previous/next subpattern
   (?R), (?0)        Recurse at the beginning of the whole pattern
   (?&name)          Recurse into a named subpattern
   (?P>name)         Recurse into a named subpattern (python syntax)
   (?(cond)yes|no)
   (?(cond)yes)      Conditional expression, where "cond" can be:
                     (N)       subpattern N has matched something
                     (<name>)  named subpattern has matched something
                     ('name')  named subpattern has matched something
                     (?{code}) code condition
                     (R)       true if recursing
                     (RN)      true if recursing into Nth subpattern
                     (R&name)  true if recursing into named subpattern
                     (DEFINE)  always false, no no-pattern allowed

=head2 VARIABLES

   $_    Default variable for operators to use

   $`    Everything prior to matched string
   $&    Entire matched string
   $'    Everything after to matched string

   ${^PREMATCH}   Everything prior to matched string
   ${^MATCH}      Entire matched string
   ${^POSTMATCH}  Everything after to matched string

The use of C<$`>, C<$&> or C<$'> will slow down B<all> regex use
within your program. Consult L<perlvar> for C<@->
to see equivalent expressions that won't cause slow down.
See also L<Devel::SawAmpersand>. Starting with Perl 5.10, you
can also use the equivalent variables C<${^PREMATCH}>, C<${^MATCH}>
and C<${^POSTMATCH}>, but for them to be defined, you have to
specify the C</p> (preserve) modifier on your regular expression.

   $1, $2 ...  hold the Xth captured expr
   $+    Last parenthesized pattern match
   $^N   Holds the most recently closed capture
   $^R   Holds the result of the last (?{...}) expr
   @-    Offsets of starts of groups. $-[0] holds start of whole match
   @+    Offsets of ends of groups. $+[0] holds end of whole match
   %+    Named capture groups
   %-    Named capture groups, as array refs

Captured groups are numbered according to their I<opening> paren.

=head2 FUNCTIONS

   lc          Lowercase a string
   lcfirst     Lowercase first char of a string
   uc          Uppercase a string
   ucfirst     Titlecase first char of a string

   pos         Return or set current match position
   quotemeta   Quote metacharacters
   reset       Reset ?pattern? status
   study       Analyze string for optimizing matching

   split       Use a regex to split a string into parts

The first four of these are like the escape sequences C<\L>, C<\l>,
C<\U>, and C<\u>.  For Titlecase, see L</Titlecase>.

=head2 TERMINOLOGY

=head3 Titlecase

Unicode concept which most often is equal to uppercase, but for
certain characters like the German "sharp s" there is a difference.

=head1 AUTHOR

Iain Truskett. Updated by the Perl 5 Porters.

This document may be distributed under the same terms as Perl itself.

=head1 SEE ALSO

=over 4

=item *

L<perlretut> for a tutorial on regular expressions.

=item *

L<perlrequick> for a rapid tutorial.

=item *

L<perlre> for more details.

=item *

L<perlvar> for details on the variables.

=item *

L<perlop> for details on the operators.

=item *

L<perlfunc> for details on the functions.

=item *

L<perlfaq6> for FAQs on regular expressions.

=item *

L<perlrebackslash> for a reference on backslash sequences.

=item *

L<perlrecharclass> for a reference on character classes.

=item *

The L<re> module to alter behaviour and aid
debugging.

=item *

L<perldebug/"Debugging regular expressions">

=item *

L<perluniintro>, L<perlunicode>, L<charnames> and L<perllocale>
for details on regexes and internationalisation.

=item *

I<Mastering Regular Expressions> by Jeffrey Friedl
(F<http://oreilly.com/catalog/9780596528126/>) for a thorough grounding and
reference on the topic.

=back

=head1 THANKS

David P.C. Wollmann,
Richard Soderberg,
Sean M. Burke,
Tom Christiansen,
Jim Cromie,
and
Jeffrey Goff
for useful advice.

=cut