Should you add HTML to the DOM using innerHTML or by creating new elements one by one?

Question

There are two methods to add HTML-code to the DOM and I don't know what is the best way to do it.

First method

The first way is the easy one, I could simply add HTML-code (with jQuery) using $('[code here]').appendTo(element); which is much like element.innerHTML = [code here];

Second method

Another way is to create all the elements one by one like:

// New div-element
var div = $('<div/>', {
              id: 'someID',
              class: 'someClassname'
          });

// New p-element that appends to the previous div-element
$('<p/>', {
    class: 'anotherClassname',
    text: 'Some textnode',
}).appendTo(div);

This method uses core functions like document.createElement and element.setAttribute.

When should I use the first method and when the second? Is method two faster than method one?

Edit - Result of speed tests

I did three speed tests from which the code follows:

$(document).ready(function(){
    // jQuery method - Above mentioned as the second method
    $('#test_one').click(function(){
        startTimer();
        var inhere = $('#inhere');
        for(i=0; i<1000; i++){
            $(inhere).append($('<p/>', {'class': 'anotherClassname' + i, text: 'number' + i}));
        }
        endTimer();
        return false;
    });

    // I thought this was much like the jQuery method, but it was not, as mentioned in the comments
    $('#test_two').click(function(){
        startTimer();
        var inhere = document.getElementById('inhere');
        for(i=0; i<1000; i++){
            var el = document.createElement('p')
            el.setAttribute('class', 'anotherClassname' + i);
            el.appendChild(document.createTextNode('number' + i));
            inhere.appendChild(el);
        }
        endTimer();
        return false;
    });

    // This is the innerHTML method
    $('#test_three').click(function(){
        startTimer();
        var inhere = document.getElementById('inhere'), el;
        for(i=0; i<1000; i++){
            el += '<p class="anotherClassname' + i + '">number' + i + '</p>';
        }                
        inhere.innerHTML = el;
        endTimer();
        return false;
    });
});

Which gave the following really surprising results

               Test One   Test Two   Test Three  
+-------------+---------+----------+------------+
| Chrome    5 | ~125ms  |  ~10ms   |   ~15ms    |
| Firefox 3.6 | ~365ms  |  ~35ms   |   ~23ms    |
| IE        8 | ~828ms  |  ~125ms  |   ~15ms    |
+-------------+---------+----------+------------+

All in all the innerHTML method seems to be the fastest one and is in many cases the most readable one.

Answer 1

I point to an outdated article for purposes of people testing this for themselves. The DOM methods actually beat out the innerHTML on all of my machines so that is what I prefer. However, at the time of the article innerHTML was faster on avg. Currently the difference can only be seen in huge datasets drastically.

Answer 2

If i am going to re-use the div later in the code, i'll build it and put it in a var, ussually with a $ prefix so i know it's a jquery object. If it's a one-off thing i'll just do a:

 $('body').append(the stuff)

Answer 3

Actually, neither methods use innerHTML, in both cases jQuery converts the HTML to DOM nodes. From jquery-1.3.2.js:

// If a single string is passed in and it's a single tag
// just do a createElement and skip the rest
if ( !fragment && elems.length === 1 && typeof elems[0] === "string" ) {
    var match = /^<(\w+)\s*\/?>$/.exec(elems[0]);
    if ( match )
        return [ context.createElement( match[1] ) ];
}

// ... some more code (shortened so nobody falls asleep) ...

// Convert html string into DOM nodes
if ( typeof elem === "string" ) {
    // Fix "XHTML"-style tags in all browsers
    elem = elem.replace(/(<(\w+)[^>]*?)\/>/g, function(all, front, tag){
        return tag.match(/^(abbr|br|col|img|input|link|meta|param|hr|area|embed)$/i) ?
            all :
            front + "></" + tag + ">";
    });
    // etc...
}

Generally speaking, using innerHTML is slower than manipulating the DOM, because it invokes the HTML parser (which will parse the HTML into the DOM anyway).