↑
Main Page

Non-capturing groups

Groups that create backreferences are called

capturing groups

. There are also

non-capturing groups

, which

don’t create backreferences. In very long regular expressions, storing backreferences slows down the

matching process. By using non-capturing groups, you can have the same flexibility to match sequences

of characters without incurring the overhead of storing the results.

If you want to create a non-capturing group, just add a question mark followed by a colon immediately

after the opening parenthesis:

var sToMatch = “#123456789”;

var reNumbers = /#(?:\d+)/;

reNumbers.test(sToMatch);

alert(RegExp.$1); //outputs “”

The last line of this example outputs an empty string because the group is specified as non-capturing.

Because of this, no backreferences can be used with the

replace()

method, accessed via the

RegExp.$x

variables, or used in the regular expression itself. Look what happens when the following code is run:

var sToMatch = “#123456789”;

var reNumbers = /#(?:\d+)/;

alert(sToMatch.replace(reNumbers, “abcd$1”)); //outputs “abcd$1”

This code outputs

“abcd$1”

instead of

“abcd123456789”

because the

“$1”

code isn’t recognized as a

backreference; instead, it is interpreted literally.

One very popular use of regular expressions is to strip HTML tags out of text. This is typically used on

discussion boards and forums to prevent visitors from including malicious or careless HTML in their

postings. To strip HTML tags using regular expressions is trivial; you just need one simple expression:

var reTag = /<(?:.|\s)*?>/g;

This expression matches a less-than symbol (

) followed by any text (specified in a non-capturing group),

followed by a greater-than symbol (

), which effectively matches all HTML tags. The non-capturing

group is used in this case because it doesn’t matter what appears between the less-than and greater-than

symbols (it all must be removed). You can use the

replace()

method with this pattern to create your

own

stripHTML()

method for a String:

String.prototype.stripHTML = function () {

var reTag = /<(?:.|\s)*?>/g;

return this.replace(reTag, “”);

};

To use this method is equally simple:

var sTest = “<b>This would be bold</b>”;

alert(sTest.stripHTML()); //outputs “This would be bold”

209

Regular Expressions

10_579088 ch07.qxd 3/28/05 11:38 AM Page 209