RegExp 物件被用來比對符合自訂規則的文字。
關於正規表達式(regular expressions)的介紹，可以閱讀 JavaScript 指南中的 正規表達式章節。
建立 RegExp 物件有兩種方式：文字表示法 (literal notation) 和 建構子 (constructor)。
以下的表達式會建立出相同的正規表達式:
/ab+c/i
new RegExp(/ab+c/, 'i') // literal notation
new RegExp('ab+c', 'i') // constructor
The literal notation provides a compilation of the regular expression when the expression is evaluated. Use literal notation when the regular expression will remain constant. For example, if you use literal notation to construct a regular expression used in a loop, the regular expression won't be recompiled on each iteration.
The constructor of the regular expression object—for example, new RegExp('ab+c')—provides runtime compilation of the regular expression. Use the constructor function when you know the regular expression pattern will be changing, or you don't know the pattern and are getting it from another source, such as user input.
Starting with ECMAScript 6, new RegExp(/ab+c/, 'i') no longer throws a {{jsxref("TypeError")}} ("can't supply flags when constructing one RegExp from another") when the first argument is a RegExp and the second flags argument is present. A new RegExp from the arguments is created instead.
When using the constructor function, the normal string escape rules (preceding special characters with \ when included in a string) are necessary.
For example, the following are equivalent:
let re = /\w+/
let re = new RegExp('\\w+')
RegExp()RegExp.prototypeRegExp.lengthRegExp.length is 2.The global RegExp object has no methods of its own. However, it does inherit some methods through the prototype chain.
RegExp prototype objects and instancesSee also deprecated RegExp properties.
Note that several of the {{JSxRef("RegExp")}} properties have both long and short (Perl-like) names. Both names always refer to the same value. Perl is the programming language from which JavaScript modeled its regular expressions.
RegExp.prototype.constructorRegExp object.. matches newlines or not.The following script uses the {{jsxref("String.prototype.replace()", "replace()")}} method of the {{jsxref("Global_Objects/String", "String")}} instance to match a name in the format first last and output it in the format last, first.
In the replacement text, the script uses $1 and $2 to indicate the results of the corresponding matching parentheses in the regular expression pattern.
let re = /(\w+)\s(\w+)/ let str = 'John Smith' let newstr = str.replace(re, '$2, $1') console.log(newstr)
This displays "Smith, John".
The default line ending varies depending on the platform (Unix, Windows, etc.). The line splitting provided in this example works on all platforms.
let text = 'Some text\nAnd some more\r\nAnd yet\rThis is the end' let lines = text.split(/\r\n|\r|\n/) console.log(lines) // logs [ 'Some text', 'And some more', 'And yet', 'This is the end' ]
Note that the order of the patterns in the regular expression matters.
let s = 'Please yes\nmake my day!' s.match(/yes.*day/); // Returns null s.match(/yes[^]*day/); // Returns ["yes\nmake my day"]
The sticky flag indicates that the regular expression performs sticky matching in the target string by attempting to match starting at {{jsxref("RegExp.prototype.lastIndex")}}.
let str = '#foo#' let regex = /foo/y regex.lastIndex = 1 regex.test(str) // true regex.lastIndex = 5 regex.test(str) // false (lastIndex is taken into account with sticky flag) regex.lastIndex // 0 (reset after match failure)
With the sticky flag y, the next match has to happen at the lastIndex position, while with the global flag g, the match can happen at the lastIndex position or later:
re = /\d/y;
while (r = re.exec("123 456")) console.log(r, "AND re.lastIndex", re.lastIndex);
// [ '1', index: 0, input: '123 456', groups: undefined ] AND re.lastIndex 1
// [ '2', index: 1, input: '123 456', groups: undefined ] AND re.lastIndex 2
// [ '3', index: 2, input: '123 456', groups: undefined ] AND re.lastIndex 3
// ... and no more match.
With the global flag g, all 6 digits would be matched, not just 3.
As mentioned above, \w or \W only matches ASCII based characters; for example, a to z, A to Z, 0 to 9, and _.
To match characters from other languages such as Cyrillic or Hebrew, use \uhhhh, where hhhh is the character's Unicode value in hexadecimal.
This example demonstrates how one can separate out Unicode characters from a word.
let text = 'Образец text на русском языке' let regex = /[\u0400-\u04FF]+/g let match = regex.exec(text) console.log(match[0]) // logs 'Образец' console.log(regex.lastIndex) // logs '7' let match2 = regex.exec(text) console.log(match2[0]) // logs 'на' [did not log 'text'] console.log(regex.lastIndex) // logs '15' // and so on
The Unicode property escapes feature introduces a solution, by allowing for a statement as simple as \p{scx=Cyrl}. One can also use an external resource for getting the complete Unicode block range for different scripts, such as Regexp-Unicode-block.
let url = 'http://xxx.domain.com' console.log(/[^.]+/.exec(url)[0].substr(7)) // logs 'xxx'
| Specification |
|---|
| {{SpecName('ESDraft', '#sec-regexp-regular-expression-objects', 'RegExp')}} |
{{Compat("javascript.builtins.RegExp")}}
Starting with Firefox 34, in the case of a capturing group with quantifiers preventing its exercise, the matched text for a capturing group is now undefined instead of an empty string:
// Firefox 33 or older
'x'.replace(/x(.)?/g, function(m, group) {
console.log("'group:" + group + "'");
});
// 'group:'
// Firefox 34 or newer
'x'.replace(/x(.)?/g, function(m, group) {
console.log("'group:" + group + "'");
});
// 'group:undefined'
Note that due to web compatibility, RegExp.$N will still return an empty string instead of undefined (bug 1053944).