regular expression

Regular expression validation tool ：https://regex101.com/

Basic characters

[] ： For custom range sets , Characters can be specified 、 Character range 、 Or fixed range set

expression	describe
[abc]	Arbitrary range character set . Match any single character contained in the set .
[^abc]	Exclude range character set . Match any character that is not in the set .
[a-z]	character in range . Match any single character in the specified range .
.	Matches any single character except the newline character , Including itself .
\	Escape character . Used to represent special characters , Such as ：^-[]$/ Equal character
\w	Match any letter 、 Numbers 、 Any single character in the underscore （ Equivalent to [A-Za-z0-9_]）.
\W	\W Express \w Other characters , Equivalent to [^A-Za-z0-9_]）.
\d	Numbers . matching 0-9 Any number .
\D	The digital . Match any non numeric characters .
\s	blank . Matches any whitespace characters , Including Spaces 、 tabs 、 Line break .
\S	Nonwhite space character . Matches any non-whitespace characters .

Multiple character sets can be combined , Form a larger character set , Such as ：[\d\s]

Logical controller

​ There are also logical evaluators in regular expressions , For example, regular Hi , Indicates that the first character is equal to H also The second character is equal to i, They are default && Calculation . If you want to express or The relationship between , Just add one in the middle | , Such as H|i Is equal to H or be equal to i.

​ There are only the following three groups of logical controllers in regular ：

Logical sign	describe
`	` or
() The subexpression shows	Used to calculate the contents in parentheses independently , Support nesting . Such as ：`( Zhang
{} Quantity control	Limit the number range of characters or word expressions . Such as ： Zhang .{1,3} Express “ Zhang ” It can only be followed by 1 To 3 Any character .

Be careful ： There is no XOR in regular ( ^ ) Logic

explain ：| In the class Unix Use problems in the system

because | In the class Unix There are special characters in the system , When you use grep or vim When performing regular search , Must be carried out escape . In the class Unix The correct posture of the system using regular search is ：

# grep  Use regular 
grep '401\|403\|404\|500' nginx.access.log

#  stay vim  Input in /  Start regular search , Again | Need to escape 
/401\|403\|404\|500

subexpression

​ Subexpression shows support nesting , Such as ：(www|mvn|test-(bj|sz|gz|sh)).coderead.cn Express test The subdomain name can be further divided into test-bj、test-sz、test-sh、test-gz.

​ The subexpression can also be used Group references .

Quantity control

quantifiers ： That is, the symbol used for range control in regular expressions , Such as * + ？ {m}

​ Quantity control acts on character or The subexpression shows , Its quantity range has been limited .\d{6} Must be 6 Digit number . Please note that ： One {} Only act on the single character before it, such as ：hi{2} Express hii, instead of hihi.

​ quantifiers * + ? In fact, it is to use {} An abbreviation for quantity control

expression	describe
?	Match the previous expression 0 Or 1 individual . Means optional . amount to {0,1}
+	Match the previous expression At least 1 individual .
*	Match the previous expression 0 One or more .
{m}	Match the previous expression m individual .
{m,}	Match the previous expression at least m individual .
{m,n}	Match the previous expression at least m individual , most n individual .

Q： How to limit the length range of text ？

A： For example, limit the length of the password to 6~20 position ：^.{6,20}$

Be careful ：? It can be used after the quantifier to express the meaning of minimizing matching （ Laziness matches ）

​ Here's an example ? Usage scenarios of , Several pairs are included below <span></span> label ？

<span>hello</span>  <span>uncle</span>

​ According to general thinking, there must be two pairs <span> label , But actually there is 3 Yes , Because for the whole text, there is another pair <span>hello...uncle</span> .

​ As shown in the figure , If you use regular expressions <span>.*<span> matching <span> label , Then it will match the largest <span>hello...uncle</span> label , Instead of matching two small ones as expected / label .

​ At this time, you can add ?, Tell the regular expression engine to match by minimization .

​ <span>.*?</span> and <span>.*</span> The difference can be understood in this way , Both regularities are based on <span> start , Anything in the middle , Until I met </span>. but .* very “ diligence ”, Even if you encounter one </span>, We will continue to look back to see if there are any , Then I found another one , Whichever is finally found . and .*? On the comparison “ lazy ”, The first one </span> Just match directly into a pair <span> label , Then go to find the next pair <span> The label .

grouping 、 Quote and replace

expression	describe
(expression)	grouping . Match the entire expression in parentheses .
(?:expression)	Non capture grouping . Match the whole string in parentheses but don't get the matching result , Can't get the group reference .
\num	A reference to the previously matched group . such as (\d)\1 You can match two identical numbers ,(Code)(Sheep)\1\2 You can match CodeSheepCodeSheep.
$ Group number	Replace . The replacement operation refers to the content to which the regularization is matched , Replace with the specified string , This string can be passed through $ Group number Reference groups for assembly . adopt $0 You can quote the whole matching content .（$0 stay Python And Javascript China does not support it. , Use $& Instead of ）

explain ：

• [] and () One difference that needs attention is [] Only single characters can be matched , and () Can match multiple characters .

[abc]  matching  a  and  b  and  c
(abc)  matching  abc

grouping

​ grouping Refers to what will match , Use ( ) Divided into multiple blocks , The divided groups can be used for extract 、 reverse quote as well as Replace operation .

give an example 1： stay javascript Use grouping in

// javascript
" Xiao Li's birthday is 2011-11-23".match(/(\d{
    4})-(\d{
    2})-(\d{
    2})/);
//  The results are as follows ：
['2011-11-23', '2011', '11', '23']
//  A pair of parentheses indicates a grouping , from 1 Start counting 

Remove group

​ () That is, it is used to express , It is also a group . If you only want to use it as a subexpression , Instead of grouping, you can use (?: ) Remove from the group list . such as (?:\d{4})-(\d{2})-(\d{2}) This expression shows that there are only two groups , month $1 And Japan $2.

Nested grouping

How is the group number named in nested groups ？ such as ： Birthday ((\d{4})-(\d{2})(\d{2})) The naming order of the group number is Open bracket The order of occurrence shall prevail .

grouping + quantifiers

​ If quantifiers are used in the same group , This group will represent multiple values , At this time through $ Group number When you extract the value, you will get the last matching value of the Group . Such as (\d)+ matching 12345, adopt $1 Will get 5

backreferences

​ backreferences ： Means to pass in expression \ Group number quote Before The grouping , But you can't quote the following content .

give an example 2： Reference grouping in regular , matching html All titles in

​ The text is as follows

<h1> First level title </h1>
<h2> Secondary title </h2>

​ If you don't use grouping , You can match as follows , But this is very cumbersome .

<h1>.*?<\/h1>|<h2>.*?<\/h2>

​ Regular expressions can be simplified by backreferencing

<(h[1-2])>.*?<\/\1>

Quote replace

​ Quote replace ： Regular has powerful substitution ability , For example, match all blank lines in the text , Then replace with empty （ Delete ）; Or match all comments , Then replace with empty ; Or match all in plain text http link , Then replace with <a> label ; Or replace the text with insert sql sentence .

give an example 3： Replace all dates below with yyyy-MM-dd Format

​ The steps to realize replacement are as follows ：

1. Write the regular matching date ：\d{4}[-.\/]\d{2}[-.\/]\d{2}
2. For date year 、 month 、 Group by day ：(\d{4})[-.\/](\d{2})[-.\/](\d{2})
3. Reference groups in replacement characters ：$1-$2-$3

​ The result of replacement

Case conversion

​ stay Idea、VS Code、Sublime、Notepad++ Wait for tools to replace , You can also use the operators in the following table for case conversion , But it is not supported in programming languages .

The operator	describe
\u Single to capital	Convert the next character to Big Write
\U Capitalize all	transformation \U After all characters turn Big Write
\U…\E Turn interval into capital	\U And \E The content of the interval is transferred Big Write
\l Single to lowercase	Convert one character to lowercase
\L All lowercase	transformation \L All characters after are converted to lowercase
\L…\E Interval to lowercase	\L And \U The contents of the interval are converted to lowercase

The specific method of use is ： Add a conversion operator to the replacement string .

give an example ： Turn the first letter of a word into capital

1. Write matching rules ：\w+
2. Capitalize and replace ：\u$0

​ The converted result is ：Hello World!

Boundary assertion

​ Boundary assertion ： Refers to a passage of expression Before and after Whether the specified Conditions , This condition consists of a sub expression , Boolean value is calculated based on whether the characters before and after match . Because the assertion part will not consume matching characters , Solid is also called Zero width assertion .

​ Boundary assertion Let regular have Conditions Ability to judge , More powerful .

Special boundary assertion

expression	describe
^	Match assertion text or line start .
$	Match assertion text or end of line .
\b	Match the beginning or end of the assertion word . such as Sheep\b Can match CodeSheep Last Sheep, Can't match CodeSheepCode Medium Sheep.
\B	Match non word boundaries or endings . such as Code\B Can match HelloCodeSheep Medium Code, Can't match HelloCode Medium Code.
\A Beginning of text	Assert the beginning of the text .JavaScript I won't support it
\Z End of text	Assert the end of the text .JavaScript I won't support it

explain ：

• ^ and $ Fit Boundary /m Use to indicate the beginning and end of each line , Otherwise, it means the beginning and end of the whole text . By m After modification ,^ Indicates the beginning of a line , Equivalent to (?<=\n|\A), By m After modification ,$ Indicates the end of the line , Equivalent to (?=\n|\Z).
• \b Word boundary is essentially a boundary assertion , Characteristics that conform to boundary assertions .\b Assert that the word is preceded by \W character , Or another boundary （ Row boundary , Text boundaries ）

\bhello\b  Equivalent to  (?<=^|\W|\A)Hello(?=$|\W|\Z)

Boundary assertion syntax

expression	describe
(?=)	Pre assertion , Determine an expression front Whether the conditions are met . such as Code(?=Sheep) Can match CodeSheep Medium Code, But can't match CodePig Medium Code.
(?!)	Prepositional negative assertion , Determine an expression front whether dissatisfaction Conditions . such as Code(?!Sheep) Can't match CodeSheep Medium Code, But it can match. CodePig Medium Code.
(?<=)	Post assertion , Determine an expression Back Whether the conditions are met . such as (?<=Code)Sheep Can match CodeSheep Medium Sheep, But can't match ReadSheep Medium Sheep.
(?<!)	Post negative assertion , Judge a paragraph of expression Back whether dissatisfaction Conditions . such as (?<!Code)Sheep Can't match CodeSheep Medium Sheep, But it can match. ReadSheep Medium Sheep.

explain ： The so-called pre post is centered on the expression , The boundary assertion is in front of the expression （ On the left ） And the back （ On the right ）

for example ：(?<= height )\d{3}(?=cm) in (?<= height ) For post assertion ( After the expression ),(?=cm) For pre assertion .

Be careful ： The post assertion is in javascript There are only Some browsers Support . In addition, post assertion is very performance consuming , So... Is not recommended infinity Quantifiers such as ：*+{n,}. Even in JAVA、Python、PHP Will report directly to Grammar mistakes .

Advanced features

1. Zero width assertion ： The assertion condition itself does not consume characters
2. ** Multi boundary assertion :** A regular rule can have multiple boundary assertions at the same time . The boundary of assertion is determined according to its location .
3. Conditional combination : Multiple sub assertions can perform Boolean operations, such as ：&&||(),
4. Any boundary ： It can be any legal expression boundary , It could even be empty Character boundary .

Conditional combination of advanced features

The boundary condition is essentially Boolean calculation , It can naturally carry out similar &&||() Boolean operation , To cope with more complex scenes . There are several ways of writing ：

1. h(?= Conditions 1)(?= Conditions 2) And operation , It means that multiple conditions must be met at the same time
2. h(?=(?= Conditions 1)(?= Conditions 2)) And operation , Conditions 1 And conditions 2 Form a new condition , And both conditions must be met
3. h(?=(?= Conditions 1)|(?! Conditions 2)) or operation , Meet the conditions 1 perhaps dissatisfaction Conditions 2
4. h(?= expression 1|(?= Conditions 1)(?! Conditions 2)) blend operation , Indicates that the expression is satisfied 1 Or meet the conditions at the same time 1、2

The above conditions are combined in Post assertion The same applies to , And there are no additional compatibility issues . In addition, any of the above conditions are based on The same boundary Calculate , That is, the characters h

Modifier

Modifier Matching logic for influencing expressions .

expression	describe
/.../i	Ignore case .（… Represents a regular expression ）
/.../g	The global matching . By default, only one result will be matched , add g All matching results will be matched
/.../m	Line boundary modifier . For multiline matching .
/.../s	After using this modifier . Can match any character , Include line breaks .
/.../x	Whitespace characters are ignored when writing regular expressions , Can write in multiple lines , And can be used # Annotate .JavaScript I won't support it

Yes /m The understanding of the ： Its function is to define ^ And $ Scope of boundary assertion . By default ^$ Represent the beginning and end of the text respectively , After adding, it means Head of line And At the end of the line , It also includes the beginning and end of the text . General regular expression applications will add m Modifier .

​ Similar in use idea,vs code, or notepad++ When waiting for tools , Its regularity has been added gmi Three modifiers , And in addition to i Don't omit case , Nothing else can be changed .

Examples of common regular expressions

details Reference resources

Reference material

• 7 Day from zero refined regular expression shows —【 Interactive learning 】
• Regular expression quick look-up table