Appendix B. Regular Expressions

The following tables summarize the regular expression grammar and syntax supported by the regular expression classes in System.Text.RegularExpressions. Each of the modifiers and qualifiers in the tables can substantially change the behavior of the matching and searching patterns. For further information on regular expressions, we recommend the definitive Mastering Regular Expressions by Jeffrey E. F. Friedl (O'Reilly).

All the syntax described in Table B-1 through Table B-10 should match the Perl5 syntax, with specific exceptions noted.

Table B-1. Character escapes

Escape Code Sequence	Meaning	Hexadecimal equivalent
\a	Bell	\u0007
\b	Backspace	\u0008
\t	Tab	\u0009
\r	Carriage return	\u000D
\v	Vertical tab	\u000B
\f	Form feed	\u000C
\n	Newline	\u000A
\e	Escape	\u001B
\040	ASCII character as octal
\x20	ASCII character as hex
\cC	ASCII control character
\u0020	Unicode character as hex
\ non-escape	A nonescape character

Special case: within a regular expression, \b means word boundary, except in a [] set, in which \b means the backspace character.

Table B-2. Substitutions

Expression	Meaning
$group-number	Substitutes last substring matched by group-number
${group-name}	Substitutes last substring matched by (?<group-name>)
$$	Substitutes a literal "$"
$&	Substitutes copy of the entire match
$'	Substitutes text of the input string preceding match
$'	Substitutes text of the input string following match
$+	Substitutes the last captures group
$_	Substitutes the entire input string

Substitutions are specified only within a replacement pattern.

Table B-3. Character sets

Expression	Meaning
.	Matches any character except \n
[characterlist]	Matches a single character in the list
[^characterlist]	Matches a single character not in the list
[char0-char1]	Matches a single character in a range
\w	Matches a word character; same as [a-zA-Z_0-9]
\W	Matches a nonword character
\s	Matches a space character; same as [ \n\r\t\v\f]
\S	Matches a nonspace character
\d	Matches a decimal digit; same as [0-9]
\D	Matches a nondigit

Table B-4. Positioning assertions

Expression	Meaning
^	Beginning of line
$	End of line
\A	Beginning of string
\Z	End of line or string
\z	Exactly the end of string
\G	Where search started
\b	On a word boundary
\B	Not on a word boundary

Table B-5. Quantifiers

Quantifier	Meaning
*	0 or more matches
+	1 or more matches
?	0 or 1 matches
{n}	Exactly n matches
{n,}	At least nmatches
{n,m}	At least n, but no more than m matches
*?	Lazy *, finds first match that has minimum repeats
+?	Lazy +, minimum repeats, but at least 1
??	Lazy ?, zero or minimum repeats
{n}?	Lazy {n}, exactly n matches
{n,}?	Lazy {n}, minimum repeats, but at least n
{n,m}?	Lazy {n,m}, minimum repeats, but at least n, and no more than m

Table B-6. Grouping constructs

Syntax	Meaning
( )	Capture matched substring
(?<name>)	Capture matched substring into group name[A]
(?<number>)	Capture matched substring into group number[A]
(?<name1-name2>)	Undefine name2 and store interval and current group into name1; if name2 is undefined, matching backtracks; name1 is optional[A]
(?: )	Noncapturing group
(?imnsx-imnsx: )	Apply or disable matching options
(?= )	Continue matching only if subexpression matches on right[C]
(?! )	Continue matching only if subexpression doesn't match on right[C]
(?<= )	Continue matching only if subexpression matches on left[B] [C]
(?<! )	Continue matching only if subexpression doesn't match on left[B][C]
(?> )	Subexpression is matched once, but isn't backtracked

^[A] Single quotes may be used instead of angle brackets, for example (?'name').

^[B] This construct doesn't backtrack; this is to remain compatible with Perl5.

^[C] Zero-width assertion; does not consume any characters.

The named capturing group syntax follows a suggestion made by Jeffrey Friedl in Mastering Regular Expressions (O'Reilly). All other grouping constructs use the Perl5 syntax.

Table B-7. Back references

Parameter syntax	Meaning
\count	Back reference count occurrences
\k<name>	Named back reference

Table B-8. Alternation

Expression syntax	Meaning
|	Logical OR
(?(expression)yes|no)	Matches yes if expression matches, else no; the no is optional
(?(name)yes|no)	Matches yes if named string has a match, else no; the no is optional

Table B-9. Miscellaneous constructs

Expression syntax	Meaning
(?imnsx-imnsx)	Set or disable options in midpattern
(?# )	Inline comment
# [to end of line]	X-mode comment (requires x option or IgnorePatternWhitespace)

Table B-10. Regular expression options

Option	RegexOption value	Meaning
i	IgnoreCase	Case-insensitive match
m	MultiLine	Multiline mode; changes ^ and $ so they match beginning and ending of any line
n	ExplicitCapture	Capture explicitly named or numbered groups
	Compiled	Compile to MSIL
s	SingleLine	Single-line mode; changes meaning of "." so it matches every character
x	IgnorePatternWhitespace	Eliminates unescaped whitespace from the pattern
	RightToLeft	Search from right to left; can't be specified in midstream