52

EDIT: WOW. This question is 12 years old now.

As someone stated, it can be done with a one-liner since 2016: https://stackoverflow.com/a/69322509/80907


The original:

I'm wondering if there's a way to change the text of anything in HTML without using innerHTML.

Reason I'm asking is because it's kinda frowned upon by the W3C. I know it's nitpicking, but I just wanna know, is there a way?

EDIT: people seem to misunderstand what I'm asking here: I want to find a way to effectivly change the text being displayed.

If I have:

<div id="one">One</a>

innerHTML allows me to do this:

var text = document.getElementsById("one");
text.innerHTML = "Two";

And the text on my screen will have changed.
I do not wish to append more text, I wish to change allready existing text.

KdgDev
  • 14,299
  • 46
  • 120
  • 156
  • 7
    FYI: innerHTML is extremly faster compared to use DOM createElement, createTextNode, etc. Efficency consideration: http://www.quirksmode.org/dom/innerhtml.html Moreover innerHTML is part of HTML5! – Marco Demaio Oct 14 '10 at 12:14
  • @Marco what do you mean when you say innerHTML is part of HTML5? – Pacerier Jun 14 '11 at 14:06
  • @Pacerier: innerHTML works in all browsers, but it is not part of the W3C HTML4.1 standards (and WebDevHobo says to be frowned upon this). Whilest innerHTML is supposed to be part of the HTML5: http://www.w3.org/TR/2011/WD-html5-diff-20110113/#htmldocument-extensions – Marco Demaio Jun 17 '11 at 14:36
  • Maybe because it's `.innerHTML` and not `.innerHtml`, Javascript (unlike PHP) is a **case sensitive** language. – Marco Demaio Jun 05 '13 at 17:14
  • @Chechulin http://www.ericvasilik.com/2006/07/code-karma.html (TL;DR: when they invented innerHTML, they couldn't make it work on tables, so they made it read-only on tables. This was fixed only in IE9. Article is by the inventor of innerHTML.) – Jasper Aug 06 '13 at 15:20

13 Answers13

47

The recommended way is through DOM manipulation, but it can be quite verbose. For example:

// <p>Hello, <b>World</b>!</p>
var para = document.createElement('p');
para.appendChild(document.createTextNode('Hello, '));

// <b>
var b = document.createElement('b');
b.appendChild(document.createTextNode('World');
para.appendChild(b);

para.appendChild(document.createTextNode('!'));

// Do something with the para element, add it to the document, etc.

EDIT

In response to your edit, in order to replace the current content, you simply remove the existing content, then use the code above to fill in new content. For example:

var someDiv = document.getElementById('someID');
var children = someDiv.childNodes;
for(var i = 0; i < children.length; i++)
    someDiv.removeChild(children[i]);

But as someone else said, I'd recommend using something like jQuery instead, as not all browsers fully support DOM, and those that do have quirks which are dealt with internally by JavaScript libraries. For example, jQuery looks something like this:

$('#someID').html("<p>Hello, <b>World</b>!</p>");
Turnor
  • 1,836
  • 12
  • 14
  • so actually, first place new text and then remove old... that seems a very lenghty way to do things... I wonder why the W3C does not approve of innerHTML – KdgDev Apr 06 '09 at 19:44
  • 2
    Other way around :) Their view is that with innerHTML, it's possible to insert invalid markup into an XML/XHTML document. Though that said, they seem to have changed their minds with HTML5. – Turnor Apr 06 '09 at 20:21
  • 15
    jQuery uses innerHTML many times inside code, therfor suggesting to use jQuery in order avoid using innerHTML does not make much sense IMHO. – Marco Demaio Oct 14 '10 at 12:13
  • Also, you should properly check your "not all browsers fully support DOM" statement in the context of this question. It's true that some browsers (mainly IE) have poor support of DOM, but, for this subject, methods are fully supported. Check compatibility tables: http://www.quirksmode.org/dom/w3c_core.html#creatingelements http://www.quirksmode.org/dom/w3c_core.html#t90 – Alejandro García Iglesias Aug 08 '12 at 14:20
  • 1
    A nicer way to remove all the childnodes would be `while (el.firstChild) { el.removeChild(el.firstChild); }` Looks cleaner and no need for an extra indexed loop – Frederik Voordeckers Jan 23 '18 at 15:21
14

The better way of doing it is to use document.createTextNode. One of the main reasons for using this function instead of innerHTML is that all HTML character escaping will be taken care of for you whereas you would have to escape your string yourself if you were simply setting innerHTML.

Andrew Hare
  • 344,730
  • 71
  • 640
  • 635
  • 5
    However, it seems that innerHTML is much faster than the W3C DOM methods: http://www.quirksmode.org/dom/innerhtml.html. – Steve Harrison Apr 05 '09 at 00:58
  • 5
    It depends what you're doing; it's misleading to say one method is blanket better. DOM tends to slow when doing childNode list operations on lots of elements (tending to O(n²)), where innerHTML can do them all at once. But innerHTML's parsing/serialisation step can make other operations slower. – bobince Apr 05 '09 at 01:53
10

You can get the same effect by manipulating the DOM. The safest way to change text is to remove all the child nodes of the element and replace them with a new text node.

var node = document.getElementById("one");

while( node.firstChild )
    node.removeChild( node.firstChild );
node.appendChild( document.createTextNode("Two") );

Removing the child nodes gets rid of the text content of your element before replacing it with the new text.

The reason most developers avoid using innerHTML is that accessing elements through the DOM is standards compliant.

Bill the Lizard
  • 398,270
  • 210
  • 566
  • 880
7

If you only want to change plain text, then there's a quicker solution that relies on standards:

document.getElementById("one").firstChild.data = "two";

Anyway, please note that innerHTML is going to be part of the upcoming HTML 5 standard.

Ionuț G. Stan
  • 176,118
  • 18
  • 189
  • 202
6

Also looking for a good alternative to bypass element.innerHTML I finally found that solution:

HTMLElement.prototype.htmlContent = function(html)
{
    var dom = new DOMParser().parseFromString('<template>'+html+'</template>', 'text/html').head;
    this.appendChild(dom.firstElementChild.content);
}

//-- document.getElementById('my-id').innerHTML = string_of_html;
document.getElementById('my-id').htmlContent(string_of_html);

Another alternative without <template> tags, but loop instead:

HTMLElement.prototype.htmlContent = function(html)
{
    var dom = new DOMParser().parseFromString(html, 'text/html').body;
    while (dom.hasChildNodes()) this.appendChild(dom.firstChild);
}

Keep in mind that this method actually 'add' content when innerHTML 'replace' content...

This may help:

HTMLElement.prototype.clearContent = function()
{
    while (this.hasChildNodes()) this.removeChild(this.lastChild);
}

//-- document.getElementById('my-id').innerHTML = '';
document.getElementById('my-id').clearContent();

doc: https://github.com/swannty/escaping-innerHTML
perf: https://jsperf.com/escaping-innerhtml

6

Simple.

Replace text.innerHTML = 'two' with text.firstChild.nodeValue = 'two'.

sth
  • 222,467
  • 53
  • 283
  • 367
BlackMagic
  • 4,378
  • 1
  • 21
  • 14
5

This is a very old question.

Now, things changed:

element.innerText = "Hi There!"

A one-liner. Since 2016 every single browser supports this.

Jongwoo Lee
  • 774
  • 7
  • 20
  • 1
    It's helpful to use when you want the content of div or p as string and not as escaped html string. – ht006 Feb 07 '23 at 07:49
3
var who=document.getElementById('one'), txt='new text';
if(who.innerText) who.innerText=txt;
else if(who.textContent) who.textContent= txt;

This may be as objectionable as innerHTML to you, but it has the advantage of working in some cases (IE) where innerHTML or appendChild do not, like some table nodes, the text of style and script elements and the value of form fields

kennebec
  • 102,654
  • 32
  • 106
  • 127
2

insertAdjacentHTML() is the way to go. Read more: Click for documentation

Heals Legodi
  • 251
  • 3
  • 6
  • While this link may answer the question, it is better to include a usable code example here and provide the link for reference. – Jeroen Heier Feb 08 '19 at 20:20
1

It appears to me that the CSS+HTML+JS combination should achieve desired effects:

.myelement:before {   
    content: attr(alt);
} 

...

<span class='myelement' alt='initial value'></span> 

...

element.setAttribute('alt', 'new value'); 

Does anyone know if this works in practice?

ilya n.
  • 18,398
  • 15
  • 71
  • 89
1

You could use DOM as follows:

<html>
<body>
<div>before</div>
<script type="text/javascript">
var element = document.getElementsByTagName("div")[0];
alert(element.firstChild.nodeValue);
element.removeChild(element.firstChild);
element.appendChild(document.createTextNode('after'));
alert(element.firstChild.nodeValue);
</script>
</body>

But I think anyone rarely do this but use a framework like jQuery or Prototype or any other javascript framework out there instead. This is jquery example:

<html>
<head>
<script type="text/javascript" src="http://ajax.googleapis.com/ajax/libs/jquery/1.2.6/jquery.min.js"></script>
</head>
<body>
<div>before</div>
<script type="text/javascript">
var element = $("div");
alert(element.text());
element.text("after");
alert(element.text());
</script>
</body>
Sergej Andrejev
  • 9,091
  • 11
  • 71
  • 108
1

Well i f i understand your question properly this should be an answer.

var text = document.getElementById("one");
//text.innerHTML = "Two";
 text.childNodes[0].nodeValue="two";
Shamik
  • 1,591
  • 2
  • 16
  • 36
0

I sometimes find it helpful to store a direct reference to the text node if I am going to be updating it regularly. I do something like this:

var dynamicText = myDiv.appendChild(document.createTextNode("the initial text"));

And then whenever I need to update it, I just do this:

dynamicText.nodeValue = "the updated text";

This prevents having to walk the DOM or add or remove any children.

Code83
  • 148
  • 2
  • 7