Help:Regular expressions

Regular expressions (shorthand regex) are patterns designed to match all or parts of a text string, often with the goal of then transforming the subject according to custom, complex logic. This is possible using metacharacters, which assign special meanings to specific keys, as pertaining to positional variables, option sets, or even groups to capture – that is, keeping "memory" of certain parts, to incorporate back into the replacement.

Usage

Parser functions

Courtesy of RegexFunctions, the {{#rmatch:}}, {{#rsplit:}}, and {{#rreplace:}} parser functions implement the most essential regex operations at the page source level. These contribute remarkable utility with other powerful extensions, particularly Cargo and DynamicPageList; One potential application might resolve parse order complexities around extension tags, to do with the infamous "QINU" issue.

Software tools

Regex enables the irreplaceable AutoWikiBrowser, an interface for performing automated, wiki-wide find-and-replace operations. Administrators have access to the on-site Special:ReplaceText tool, a less robust but still highly convenient alternative.

Note also that JavaScript supports regex; Lua uses its own system, though they are quite similar.

Resources

Metacharacter cheat sheet

#	Key	Name	Meaning
1.	\	Backslash	Escape a metacharacter to its literal meaning, -or- signify a token.
2.	^	Caret	Anchor start of string.
3.	|	Pipe	Logical OR: Designate alternation between options.
4.	.	Period	Wildcard: Match any character except newline.
5.	$	Dollar sign	Anchor end of string.
6.	?	Question mark	Quantifier: Zero or one i.e. "Optional"; See also greediness.
7.	*	Asterisk	Quantifier: Zero or more
8.	+	Plus sign	Quantifier: One or more
9.	(	Opening Parenthesis	Isolate a capture group. Reference contents in output via positional index $1, $2, etc.
10.	)	Closing Parenthesis
11.	[	Opening Square bracket	Define a character class: option list or range of character selection. ] doesn't normally need escape as meaning depends on the former.
12.	{	Opening curly bracket	Quantifier: Enclosed n (any integer). Optionally add , with or without upper limit p (p > n). Ending note above applies also to }.

Tutorials

• Regular expression on Wikipedia
• AutoWikiBrowser pattern guide
• Regular expressions - JavaScript flavor
• Regular expression HOWTO - Python flavor
• Regular-Expressions.info
• www.regextutorial.org

Important to familiarize is Multiline Mode, in which the anchor keys ^ and $ match beginning and end of the entire string, not one single line as standard.

Convenient examples

Pattern	Application	Tests
\[\[([^|\]\n]+)(?:\|([^\]\n]+))?\]\]	Extract the target page $1 and if applicable, display text $2 of an internal link. Replace each [/] with {/} to match an equivalent template call.	Input: [[Page Link|Page Text]]	Capture return: $1 = Page Link $2 = Page Text
<([a-z]+)>(.*?)<\/\1>	Get the name $1 and contents $2 of an HTML tag set.	Input: <span>Example text</span>	Capture return: $1 = span $2 = Example text
^(.*?)\:(.*?)	Capture prefix $1 and page name $2 from a full page title.	Input: Help:Regular expressions	Capture return: $1 = Help $2 = Regular expressions
(\x7F)*\'\"\`UNIQ(.*?)QINU\`\"\'(\x7F)*	Eliminate failed strip marker substitutions from the output, keeping their inner content $1.	Input: '"`UNIQ--ref-00000023-QINU`"'	Capture return: $1 = --ref-00000023-