jQuery document.createElement equivalent?

Rob Stevenson-Leggett picture Rob Stevenson-Leggett · Nov 6, 2008 · Viewed 924.1k times · Source

I'm refactoring some old JavaScript code and there's a lot of DOM manipulation going on.

var d = document;
var odv = d.createElement("div");
odv.style.display = "none";
this.OuterDiv = odv;

var t = d.createElement("table");
t.cellSpacing = 0;
t.className = "text";
odv.appendChild(t);

I would like to know if there is a better way to do this using jQuery. I've been experimenting with:

var odv = $.create("div");
$.append(odv);
// And many more

But I'm not sure if this is any better.

Answer

nickf picture nickf · Nov 6, 2008

Here's your example in the "one" line.

this.$OuterDiv = $('<div></div>')
    .hide()
    .append($('<table></table>')
        .attr({ cellSpacing : 0 })
        .addClass("text")
    )
;

Update: I thought I'd update this post since it still gets quite a bit of traffic. In the comments below there's some discussion about $("<div>") vs $("<div></div>") vs $(document.createElement('div')) as a way of creating new elements, and which is "best".

I put together a small benchmark, and here are roughly the results of repeating the above options 100,000 times:

jQuery 1.4, 1.5, 1.6

               Chrome 11  Firefox 4   IE9
<div>            440ms      640ms    460ms
<div></div>      420ms      650ms    480ms
createElement    100ms      180ms    300ms

jQuery 1.3

                Chrome 11
<div>             770ms
<div></div>      3800ms
createElement     100ms

jQuery 1.2

                Chrome 11
<div>            3500ms
<div></div>      3500ms
createElement     100ms

I think it's no big surprise, but document.createElement is the fastest method. Of course, before you go off and start refactoring your entire codebase, remember that the differences we're talking about here (in all but the archaic versions of jQuery) equate to about an extra 3 milliseconds per thousand elements.


Update 2

Updated for jQuery 1.7.2 and put the benchmark on JSBen.ch which is probably a bit more scientific than my primitive benchmarks, plus it can be crowdsourced now!

http://jsben.ch/#/ARUtz