So this is an interesting one and may very well be impossible to do efficiently. But I'm interested in finding an efficient way to query all html elements in the document that have a particular value set for any attribute. So, for example, instead of this:
document.querySelectorAll('[attrName]');
I'm looking for the equivalent of the following pseudo-code:
document.querySelectorAll('[*=specificValue]');
So essentially the result would be all elements that have any attribute whose value matches "specificValue". I know there are gross ways to do this such as:
var all = document.querySelectorAll('*');
var matches = [];
var attrs;
for (var i = 0; i < all.length; i += 1) {
attrs = Array.prototype.slice.call(all[i].attributes);
for (var j = 0; j < attrs.length; j += 1) {
if (attrs[j].value === 'specificValue') {
matches.push(all[i]);
break;
}
}
}
However, I would really love to avoid analyzing every single html element like this. Any ideas?
Edit:
Thanks for all the help so far. Before too many people give alternate suggestions I should explain what this is for. Basically, it's an experiment. The idea was that I might be able to create live object-to-dom databindings like what you get in Ember.js but instead of having to compile templates, you could just use regular html attributes and have a syntax marker in the value itself that a binding should be created. For example: <a href="{{linkLocation}}"></a>. I figured this might be fun if there was an efficient way to select relevant elements on the fly. Clearly I know a loop has to happen somewhere. However, if the browser is running a native iteration, I'd prefer that over my own JavaScript loop. I just wasn't sure if there was some "secret" selector syntax I wasn't aware of or if anyone could think of any other cool tricks.
href,id, orclass, for instance?