The [@@match]() method of a regular expression specifies how String.prototype.match() should behave. In addition, its presence (or absence) can influence whether an object is regarded as a regular expression.

Try it





A String that is a target of the match.

Return value

An Array whose contents depend on the presence or absence of the global (g) flag, or null if no matches are found.

  • If the g flag is used, all results matching the complete regular expression will be returned, but capturing groups are not included.
  • If the g flag is not used, only the first complete match and its related capturing groups are returned. In this case, match() will return the same result as RegExp.prototype.exec() (an array with some extra properties).


This method is called internally in String.prototype.match().

For example, the following two examples return same result.



If the regex is global (with the g flag), the regex's exec() method will be repeatedly called until exec() returns null. Otherwise, exec() would only be called once and its result becomes the return value of @@match.

Because @@match would keep calling exec() until it returns null, and exec() would automatically reset the regex's lastIndex to 0 when the last match fails, @@match would typically not have side effects when it exits. However, when the regex is sticky but not global, lastIndex would not be reset. In this case, each call to match() may return a different result.

const re = /[abc]/y;
for (let i = 0; i < 5; i++) {
  console.log("abc".match(re), re.lastIndex);
// [ 'a' ] 1
// [ 'b' ] 2
// [ 'c' ] 3
// null 0
// [ 'a' ] 1

When the regex is sticky and global, it would still perform sticky matches — i.e. it would fail to match any occurrences beyond the lastIndex.

console.log("ab-c".match(/[abc]/gy)); // [ 'a', 'b' ]

If the current match is an empty string, the lastIndex would still be advanced — if the regex has the u flag, it would advance by one Unicode codepoint; otherwise, it advances by one UTF-16 code unit.

console.log("😄".match(/(?:)/g)); // [ '', '', '' ]
console.log("😄".match(/(?:)/gu)); // [ '', '' ]

This method exists for customizing match behavior within RegExp subclasses.

In addition, the @@match property is used to check whether an object is a regular expression — only when it's undefined will the language fall back to a branded check of whether the object actually extends RegExp.prototype. For an example, see Symbol.match.


Direct call

This method can be used in almost the same way as String.prototype.match(), except the different this and the different arguments order.

const re = /[0-9]+/g;
const str = '2016-01-02';
const result = re[Symbol.match](str);
console.log(result);  // ["2016", "01", "02"]

Using @@match in subclasses

Subclasses of RegExp can override the [@@match]() method to modify the default behavior.

class MyRegExp extends RegExp {
  [Symbol.match](str) {
    const result = RegExp.prototype[Symbol.match].call(this, str);
    if (!result) return null;
    return {
      group(n) {
        return result[n];

const re = new MyRegExp('([0-9]+)-([0-9]+)-([0-9]+)');
const str = '2016-01-02';
const result = str.match(re); // String.prototype.match calls re[@@match].
console.log(; // 2016
console.log(; // 01
console.log(; // 02


ECMAScript Language Specification
# sec-regexp.prototype-@@match

Browser compatibility

BCD tables only load in the browser

See also