C Regex Cheat Sheet

C# Regex Cheat Sheet
Regex Cheat Sheet Powershell
C Sharp Regex Cheat Sheet
Regular Expression Cheat Sheet Python

characters — what to seek

C# Regex Cheat Sheet

OverAPI.com is a site collecting all the cheatsheets,all! The POSIX-Extended regular expression syntax is supported by the POSIX C regular expression API's, and variations are used by the utilities egrep and awk. You can construct POSIX extended regular expressions in Boost.Regex by passing the flag extended to the regex constructor, for example.

ring matches ring, springboard, ringtone, etc.

. matches almost any character

h.o matches hoo, h2o, h/o, etc.

Use to search for these special characters:

[ ^ $ . | ? * + ( ) { }

ring? matches ring?

(quiet) matches (quiet)

c:windows matches c:windows

alternatives — | (OR)

cat|dog match cat or dog

order matters if short alternative is part of longer

id|identity matches id or identity

regex engine is 'eager', stops comparing
as soon as 1st alternative matches

identity|id matches id or identity

order longer to shorter when alternatives overlap

(To match whole words, see scope and groups.)

character classes — [allowed] or [^NOT allowed]

[aeiou] match any vowel

[^aeiou] match a NON vowel

r[iau]ng match ring, wrangle, sprung, etc.

gr[ae]y match gray or grey

[a-zA-Z0-9] match any letter or digit

(In [ ] always escape . ] and sometimes ^ - .)

shorthand classes

w 'word' character (letter, digit, or underscore)

d digit

s whitespace (space, tab, vtab, newline)

W, D, or S, (NOT word, digit, or whitespace)

[DS] means not digit OR whitespace, both match

[^ds] disallow digit AND whitespace

occurrences — ? * + {n} {n,} {n,n}

? 0 or 1

colou?r match color or colour

* 0 or more

[BW]ill[ieamy's]* match Bill, Willy, William's etc.

+ 1 or more

[a-zA-Z]+ match 1 or more letters

{n} require n occurrences

d{3}-d{2}-d{4} match a SSN

{n,} require n or more

[a-zA-Z]{2,} 2 or more letters

{n,m} require n - m

[a-z]w{1,7} match a UW NetID

* greedy versus *? lazy

* + and {n,} are greedy — match as much as possible

<.+> finds 1 big match in bold

*? +? and {n,}? are lazy — match as little as possible

<.+?> finds 2 matches in bold

comments — (?#comment)

(?#year)(19|20)dd embedded comment

(?x)(19|20)dd #year free spacing & EOL comment

(see modifiers)

scope — b B ^ $

b 'word' edge (next to non 'word' character)

bring word starts with 'ring', ex ringtone

ringb word ends with 'ring', ex spring

b9b match single digit 9, not 19, 91, 99, etc..

b[a-zA-Z]{6}b match 6-letter words

B NOT word edge

BringB match springs and wringer

^ start of string $ end of string

^d*$ entire string must be digits

^[a-zA-Z]{4,20}$ string must have 4-20 letters

^[A-Z] string must begin with capital letter

[.!?')]$ string must end with terminal puncutation

groups — ( )

(in|out)put match input or output

d{5}(-d{4})? US zip code ('+ 4' optional)

Locate all PHP input variables:

NB: parser tries EACH alternative if match fails after group.
Can lead to catastrophic backtracking.

back references — n

each ( ) creates a numbered 'back reference'

(to) (be) or not 1 2 match to be or not to be

([^s])1{2} match non-space, then same twice more aaa, ...

b(w+)s+1b match doubled words

non-capturing group — (?: ) prevent back reference

on(?:click|load) is faster than on(click|load)

use non-capturing or atomic groups when possible

atomic groups — (?>a|b) (no capture, no backtrack)

(?>red|green|blue)

faster than non-capturing

alternatives parsed left to right without return

(?>id|identity)b matches id, but not identity

'id' matches, but 'b' fails after atomic group,
parser doesn't backtrack into group to retry 'identity'

If alternatives overlap, order longer to shorter.

lookahead — (?= ) (?! ) lookbehind — (?<= ) (?<! )

bw+?(?=ingb) match warbling, string, fishing, ...

b(?!w+ingb)w+b words NOT ending in 'ing'

(?<=bpre).*?b match pretend, present, prefix, ...

bw{3}(?<!pre)w*?b words NOT starting with 'pre'

(lookbehind needs 3 chars, w{3}, to compare w/'pre')

bw+(?<!ing)b match words NOT ending in 'ing'

(see LOOKAROUND notes below)

if-then-else — (?ifthen|else)

match 'Mr.' or 'Ms.' if word 'her' is later in string

M(?(?=.*?bherb)s|r). lookahead for word 'her'

Regex Cheat Sheet Powershell

(requires lookaround for IF condition)

modifiers — i s m x

ignore case, single-line, multi-line, free spacing

(?i)[a-z]*(?-i) ignore case ON / OFF

(?s).*(?-s) match multiple lines (causes . to match newline)

(?m)^.*;$(?-m)^ & $ match lines not whole string

(?x) #free-spacing mode, this EOL comment ignored

d{3} #3 digits (new line but same pattern)

-d{4} #literal hyphen, then 4 digits

(?-x) (?#free-spacing mode OFF)

/regex/ismx modify mode for entire string

A few examples:

(?s)<p(?(?=s) .*?)>(.*?) span multiple lines
(?s)<p(?(?=s) .*?)>(.*?) locate opening '<p'
(?s)<p(?(?=s) .*?)>(.*?) create an if-then-else
(?s)<p(?(?=s) .*?)>(.*?) lookahead for a whitespace character
(?s)<p(?(?=s) .*?)>(.*?) if found, attempt lazy match of any characters until ...
(?s)<p(?(?=s) .*?)>(.*?) closing angle brace
(?s)<p(?(?=s) .*?)>(.*?) capture lazy match of all characters until ...
(?s)<p(?(?=s) .*?)>(.*?) closing ''

The lookahead prevents matches on PRE, PARAM, and PROGRESS tags by only allowing more characters in the opening tag if P is followed by whitespace. Otherwise, '>' must follow '<p'.

LOOKAROUND notes

(?= ) if you can find ahead
(?! ) if you can NOT find ahead
(?<= ) if you can find behind
(?<! ) if you can NOT find behind

convert Firstname Lastname to Lastname, Firstname (& visa versa)

Pattern below uses lookahead to capture everything up to a space, characters, and a newline.
The 2nd capture group collects the characters between the space and the newline.
This allows for any number of names/initials prior to lastname, provided lastname is at the end of the line.

Find: (.*)(?= .*n) (.*)n

Repl: 2, 1n — insert 2nd capture (lastname) in front of first capture (all preceding names/initials)

Reverse the conversion.

C Sharp Regex Cheat Sheet

Find: (.*?), (.*?)n — group 1 gets everything up to ', ' — group 2 gets everything after ', '

Repl: 2 1n

lookaround groups are non-capturing

If you need to capture the characters that match the lookaround condition, you can insert a capture group inside the lookaround.

(?=(sometext)) the inner () captures the lookahead

This would NOT work: ((?=sometext)) Because lookaround groups are zero-width, the outer () capture nothing.

lookaround groups are zero-width

They establish a condition for a match, but are not part of it.

Compare these patterns: re?d vs r(?=e)d

re?d — match an 'r', an optional 'e', then 'd' — matches red or rd

r(?=e)d — match 'r' (IF FOLLOWED BY 'e') then see if 'd' comes after 'r'

The lookahead seeks 'e' only for the sake of matching 'r'.
Because the lookahead condition is ZERO-width, the expression is logically impossible.
It requires the 2nd character to be both 'e' and 'd'.
For looking ahead, 'e' must follow 'r'.
For matching, 'd' must follow 'r'.

fixed-width lookbehind

Most regex engines depend on knowing the width of lookbehind patterns. Ex: (?<=h1) or (?<=w{4}) look behind for 'h1' or for 4 'word' characters.

This limits lookbehind patterns when matching HTML tags, since the width of tag names and their potential attributes can't be known in advance.

variable-width lookbehind

.NET and JGSoft support variable-width lookbehind patterns. Ex: (?<=w+) look behind for 1 or more word characters.
The first few examples below rely on this ability.

Lookaround groups define the context for a match. Here, we're seeking .* ie., 0 or more characters.
A positive lookbehind group (?<= . . . ) preceeds. A positive lookahead group (?= . . . ) follows.
These set the boundaries of the match this way:

(?<=<(w+)>).*(?=</1>) look behind current location
(?<=<(w+)>).*(?=</1>) for < > surrounding ...
(?<=<(w+)>).*(?=</1>) one or more 'word' characters. The ( ) create a capture group to preserve the name of the presumed tag: DIV, H1, P, A, etc.
(?<=<(w+)>).*(?=</1>) match anything until
(?<=<(w+)>).*(?=</1>) looking ahead from the current character
(?<=<(w+)>).*(?=</1>) these characters surround
(?<=<(w+)>).*(?=</1>) the contents of the first capture group

In other words, advance along string until an opening HTML tag preceeds. Match chars until its closing HTML tag follows.
The tags themselves are not matched, only the text between them.

Regular Expression Cheat Sheet Python

To span multiple lines, use the (?s) modifier. (?s)(?<=<cite>).*(?=</cite>) Match <cite> tag contents, regardless of line breaks.

As in example above, the first group (w+) captures the presumed tag name, then an optional space and other characters ?.*? allow for attributes before the closing >.

class='.*?bredb.*?' this new part looks for class=' and red and ' somewhere in the opening tag
b ensures 'red' is a single word
.*? allow for other characters on either side of 'red' so pattern matches class='red' and class='blue red green' etc.

Here, the first group captures only the tag name. The tag's potential attributes are outside the group.

(?i)<([a-z][a-z0-9]*)[^>]*>.*?</1> set ignore case ON
(?i)<([a-z][a-z0-9]*)[^>]*>.*?</1> find an opening tag by matching 1 letter after <
(?i)<([a-z][a-z0-9]*)[^>]*>.*?</1> then match 0 or more letters or digits
(?i)<([a-z][a-z0-9]*)[^>]*>.*?</1> make this tag a capture group
(?i)<([a-z][a-z0-9]*)[^>]*>.*?</1> match 0 or more characters that aren't > — this allows attributes in opening tag
(?i)<([a-z][a-z0-9]*)[^>]*>.*?</1> match the presumed end of the opening tag
(NB: This markup <a> would end the match early. Doesn't matter here. Subsequent < pulls match to closing tag. But if you attempted to match only the opening tag, it might be truncated in rare cases.)
(?i)<([a-z][a-z0-9]*)[^>]*>.*?</1> lazy match of all of tag's contents
(?i)<([a-z][a-z0-9]*)[^>]*>.*?</1> match the closing tag — 1 refers to first capture group

The IF condition can be set by a backreference (as here) or by a lookaround group.

(()?d{3} optional group ( )? matches '(' prior to 3-digit area code d{3} — group creates back reference #1

(?(1)) ?|[-/ .]) (1) refers to group 1, so if '(' exists, match ')' followed by optional space, else match one of these: '- / . '

d{3}[- .]d{4} rest of phone number

For a quick overview: http://www.codeproject.com/KB/dotnet/regextutorial.aspx.

For a good tutorial: http://www.regular-expressions.info.

Anchors

Anchros

^
Start of string, or start of line in multi-line pattern
A
Start of string
$
End of string, or end of line in multi-line pattern
Z
End of string
b
Word boundary
B
Not word boundary
<
Start of word
>
End of word

Quantifiers

*
0 or more
+
1 or more
?
0 or 1
{3}
Exactly 3
{3,}
3 or more
{3,5}
3, 4 or 5
{,5}
5 or less

Character

Character Classes

c
Control character
s
White space
S
Not white space
d
Digit
D
Not digit
w
Word
W
Not word
x
Hexade-cimal digit
O
Octal digit

Special

n
New line
r
Carriage return
t
Tab
v
Vertical tab
f
Form feed
xxx
Octal character xxx
xhh
Hex character hh

Examples

Metacharacter

^abc
abc, abcdefg, abc123, ...
abc$
abc, endsinabc, 123abc, ...
a.c
abc, aac, acc, adc, aec, ...
bill|ted
ted, bill
ab{2}c
abbc
a[bB]c
abc, aBc
(abc){2}
abcabc
ab*c
ac, abc, abbc, abbbc, ...
ab+c
abc, abbc, abbbc, ...
ab?c
ac, abc
asc
a c

Sample

([A-Za-z0-9-]+)
Letters, numbers and hyphens
(d{1,2}/d{1,2}/d{4})
Date (e.g. 21/3/2006)
([^s]+(?=.(jpg|gif|png)).2)
jpg, gif or png image
(^[1-9]{1}$|^[1-4]{1}[0-9]{1}$|^50$)
Any number from 1 to 50 inclusive
(#?([A-Fa-f0-9]){3}(([A-Fa-f0-9]){3})?)
Valid hexadecimal colour code
((?=.*d)(?=.*[a-z])(?=.*[A-Z]).{8,15})
8 to 15 character string with at least one upper case letter, one lower case letter, and one digit (useful for passwords).
(w+@[a-zA-Z_]+?.[a-zA-Z]{2,6})
Email addresses
(<(/?[^>]+)>)
HTML Tag

POSIX

[:upper:]
Upper case letters
[:lower:]
Lower case letters
[:alpha:]
All letters
[:alnum:]
Digits and letters
[:digit:]
Digits
[:xdigit:]
Hexade-cimal digits
[:punct:]
Punctu-ation
[:blank:]
Space and tab
[:space:]
Blank characters
[:cntrl:]
Control characters
[:graph:]
Printed characters
[:print:]
Printed characters and spaces
[:word:]
Digits, letters and underscore

Groups

Groups and Ranges

.
Any character except new line (n)
(a|b)
a or b
(...)
Group
(?:...)
Passive (non-c-apt-uring) group
[abc]
Range (a or b or c)
[^abc]
Not a or b or c
[a-q]
Letter from a to q
[A-Q]
Upper case letter from A to Q
[0-7]
Digit from 0 to 7
n
nth group/-sub-pattern

Modifiers

g
Global match
i
Case-i-nse-nsitive
m
Multiple lines
s
Treat string as single line
x
Allow comments and white space in pattern
e
Evaluate replac-ement
U
Ungreedy pattern

String

Replacement

$n
nth non-pa-ssive group
$2
'-xyz-' in /^(abc-(xy-z))$/
$1
'-xyz-' in /^(?:a-bc)-(xyz)$/
$`
Before matched string
$'
After matched string
$+
Last matched string
$&
Entire matched string

Assertions

?=
Lookahead assertion
?!
Negative lookahead
?<=
Lookbehind assertion
?!= or ?<!
Negative lookbehind
?>
Once-only Subexp-ression
?()
Condition [if then]
?()|
Condition [if then else]
?#
Comment