views:

298

answers:

3

I got to thinking today: what is the best way of getting a distinct (ie no repeats) list of classes used in a document that (preferably) match a pattern (regular expression) pattern or (alternatively) start with a certain character sequence? JQuery can be used for this or just straight Javascript.

Now it should obviously cater for all legal class usages, for example:

<div class="class1 class2 class3">
</div>

And I don't want to parse the document with regular expressions. That's simply too error prone. What I'm interested in is a Jaavascript solution that walks the DOM or uses something like jQuery to do that.

Oh this should also include any classes that have been dynamically added/removed through previous Javascript code.

Suggestions?

+1  A: 

Can this help you?

http://httpcode.com/blogs/PermaLink,guid,77f17d9c-cdc0-4fcb-a392-3cc70ef6ac23.aspx

<input type="radio" class="star star_id_45 star_group_5" />

$("input[class*='star_id_45']")
Natrium
Sort of. I didn't realize you could do the wildcard matching with the class attribute (seen it for href) but I guess there's no reason why not. Just hadn't occurred to me. But I'm really after the distinct class names rather than a wrapped set of elements with them.
cletus
+2  A: 
function gatherClasses() {
    var tags = document.getElementsByTagName('*');
    var cls, clct = {}, i, j, l = tags.length;
    for( i = 0; i < l; i++ ) {
        cls = tags[i].className.split(' ');
        for( j = 0; j < cls.length; j++ ) {
            if( !cls[j] ) continue;
            clct[cls[j]] = 'dummy'; //so we only get a class once
        }
    }
    cls = [];
    for( i in clct ) {
        cls.push( i );
    }
    return cls;
} 
alert(gatherClasses())

Heres a version with a regexp match

function gatherClasses( matchString ) {
    if( matchString ) {
        var rxp = new RegExp( matchString );
    } else {
        var rxp = /.*/;
    }
    var tags = document.getElementsByTagName('*');
    var cls, clct = {}, i, j, l = tags.length;
    for( i = 0; i < l; i++ ) {
        cls = tags[i].className.split(' ');
        for( j = 0; j < cls.length; j++ ) {
            if( !cls[j] || !rxp.test( cls[j] ) ) {
                continue;
            }
            clct[cls[j]] = 'dummy'; //so we only get a class once
        }
    }
    cls = [];
    for( i in clct ) {
        cls.push( i );
    }
    return cls;
}
//find classes that match 'stack'
alert(gatherClasses('stack'))
meouw
This works too but the other solution is just a little terser (good thing) and a little more robust (eg splitting on \s+ rather than ' ').
cletus
+1  A: 

Using jQuery:

var listClasses = function( pattern ){
    var allClassesTmp = {}, allClasses = []; 
    var rx = pattern ? (new RegExp(pattern)) : (new RegExp(".*"));

    $('*[class]').each( function(){ 
        var cn = this.className.split(/\s+/); 
        for(var i=cn.length;--i>-1;){ 
            if(rx.test(cn[i]))allClassesTmp[ cn[i] ] = 1 
        } 
    }); 
    for(var i in allClassesTmp)allClasses.push(i); 
    return allClasses;
}
pawel