Develop Reference

looping over array of elements just affects the last element

In one of my projects I just discovered, that sometimes iterating over an array of html elements (and change all of them) just affects the last element. When I log the element's attributes I can see that the loop definitily adresses every element but nevertheless visibly just the last element is getting changed.
Can anyone explain me why?
I already figured out, that a solution is to use createElement() and appendChild() instead of insertHTML. I just want to understand why javascript behaves like this.
Here is my example code:
/* creating 5 elements and storing them into an array */
var elementArray = [];
for(var n = 0;n<5;n++)
{
document.body.innerHTML += "<div id='elmt_"+n+"'>"+n+"</div>\n";
elementArray[n] = document.getElementById("elmt_"+n);
}
/* loop over these 5 elements */
for(var n = 0;n<5;n++)
{
console.log(elementArray[n].id); // logs: elmt_0 elmt_1 elmt_2 elmt_3 elmt_4
elementArray[n].innerHTML = "test"; // changes just the last element (elmt_4) to "test"
}
I created an example here: http://jsfiddle.net/qwe44m1o/1/

1 - Using console.log(elementArray[n]); in your second loop shows that innerHTML in this loop is modifying html inside your array, not in your document. That means that you are storing the div element in your array, not a shortcut to document.getElementById("elmt_"+n)
See the JSFiddle
2 - If you want to store a shortcut in order to target an element by ID, you have to add quotes for elementArray[n] = "document.getElementById('elmt_"+n+"')";, and use it with eval like this : eval(elementArray[n]).innerHTML = n+"-test";
See the JSFiddle for this try

Finding the index of the element with class in native Javascript

Is there a way to get the index of class name (I.e. the third element with the class "className" would be 3 without using jQ?
I don't know jQ, and I don't have time to learn it right now, and I don't want to include code into my code that I don't understand at least some.
Thanks.
BTW, I've used jQ instead of spelling it out so those results can be filtered out in Google should somebody have the same question. If I spelled it out, and somebody used the NOT operator in Google, this one would also disappear.

You could do something like:
// the element you're looking for
var target = document.getElementById("an-element");
// the collection you're looking in
var nodes = document.querySelectorAll(".yourclass");
var index = [].indexOf.call(nodes, target);
See: Array's indexOf.
If you have already a proper array as nodes instead of a NodeList, you can just do nodes.indexOf(target).

you can use document.getElementsByClassName
var el = document.getElementsByClassName('className');
for (var i = 0; i < el.length; i++) {
// your index is inside here
}
el[i] is the element in the current iteration, i is the index
(I.e. the third element with the class "className" would be 3)
if (i == 3)
return el[i]
JsFiddle: here.

Just use getElementsByClassName, it returns a list of elements with the specified classes.
elements = document.getElementsByClassName("test")
element = elements[2] //get the 3rd element
Hope this helps!

these work as of es6:
Array.from(document.querySelectorAll('.elements')).indexOf(anElement)
or
[...document.querySelectorAll('.elements')].indexOf(anElement)

Why is this while loop infinite? JavaScript appendChild

I swear this was just working fine a few days ago...
elm = document.querySelectorAll(selector);
var frag = document.createDocumentFragment();
while (elm[0]){
frag.appendChild(elm[0]);
}
Right, so, this should append each node from our elm node list. When the first one is appended, the second "moves" to the first position in node list, hence the next one is always elm[0]. It should stop when the elm nodeList is completely appended. However, this is giving me an infinite loop. Thoughts?
EDIT - because I've gotten the same answer several times...
A nodeList is not an array, it is a live reference. When a node is "moved" (here, appended) it should be removed automatically from the node list. The answers all saying "you're appending the same element over and over" - this is what's happening, it shouldn't be. A for loop shouldn't work, because when the first node is appended, the next node takes its index.
2nd EDIT
So the question is now "why is the nodeList behaving as an array?". The node list SHOULD be updating every time a node is being appended somewhere. Most peculiar.
Solution (in case someone needs something to handle live + non-live node lists)
elm = (/*however you're getting a node list*/);
var frag = document.createDocumentFragment();
var elength = elm.length;
for (var b = 0; b<elength; b++){
if (elm.length === elength){
frag.appendChild(elm[b]);
} else {
frag.appendChild(elm[0].cloneNode());
}
}
Basically, just checking to see if the node list has changed length.

From the MDN Docs
Element.querySelectorAll
Summary
Returns a non-live NodeList of all elements descended from the element on which it is invoked that match the specified group of CSS selectors.
Syntax
elementList = baseElement.querySelectorAll(selectors);
where
elementList is a non-live list of element objects.
baseElement is an element object.
selectors is a group of selectors to match on.
From the docs above you can see it does not automatically remove it when you append it to another element since it is non live. Run a demo to show that feature.
var selector = "div";
elm = document.querySelectorAll(selector);
var frag = document.createDocumentFragment();
console.log("before",elm.length);
frag.appendChild(elm[0]);
console.log("after",elm.length);
When the code above runs, in the console you get.
before 3
after 3
If you want to do the while loop, convert to an array and shift() the items off
var selector = "div";
var elmNodeLIst = document.querySelectorAll(selector);
var frag = document.createDocumentFragment();
var elems = Array.prototype.slice.call(elmNodeLIst );
while (elems.length) {
frag.appendChild(elems.shift());
}
console.log(frag);

You are appending the first item in the node list, over and over and over. You never removing any items from the array, but always adding the first one to the fragment. And the first one is always the same.
elm = document.querySelectorAll(selector);
var frag = document.createDocumentFragment();
while (elm.length){
frag.appendChild(elm.shift());
}
This may be closer to what you meant to do. We can use while (elm.length) because as items get removed form the array, eventually length will be zero which is a flasy value and the loop will stop.
And we use elm.shift() to fetch the item from the array because that method will return the item at index zero and remove it from the array, which gives us the mutation of the original array we need.
I think you thought this might work because a node can only have one parent. Meaning adding somewhere removes it from the previous parent. However, elm is not a DOM fragment. It's just a aray (or perhaps a NodeList) that holds references to element. The array is not the parent node of these elements, it just holds references.
Your loop might work if you had it like this, since you are query the parent node each time for its children, a list of node that will actually change as you move around:
elm = document.getElementById(id);
var frag = document.createDocumentFragment();
while (elm.children[0]){
frag.appendChild(elm.children[0]);
}

I wouldn't have expected it to work in the first place.
Your elm array is initialized, and never updated. Even if the result from running document.querySelectorAll(selector); would return something different, this doesn't change your current references in the array.
You would either need to rerun the selector, or manually remove the first element in the array after appending it.

elm[0] is static and unchanging in above code
fix is as below
elm = document.querySelectorAll(".container");
var frag = document.createDocumentFragment();
console.log(elm);
var i=0;
while (elm[i]){
frag.appendChild(elm[i++]);
}

I didn't actually focus much on the code (and if it made sense -judging from the comments- or not); but if this worked a few days ago then the problem is in the input you are giving to your code selector.
That's when Unit Testing comes in handy. If you can remember the input with which the code worked, then you can make it work again and start debugging from there.
Otherwise, you are just lying to yourself.

It's an infinite loop as it's written right now because elm[0] always refers to the same element, and that element is never null (any non-null/non-zero result would be true). You also don't do anything with the elements themselves to make it iterate across the list. You should be using a for loop instead of a while or at least having some kind of indexer to try to traverse the collection.
elm = document.querySelectorAll(selector);
var frag = document.createDocumentFragment();
for (i= 0; i < elm.length; i++)
{
frag.appendChild(elm[i]);
}
Edit:
From the documentation:
A "live" collection
In most cases, the NodeList is a live collection. This means that changes on the DOM tree >are going to be reflected on the collection.
var links = document.getElementsByTagName('a'); // links.length === 2 for instance
document.body.appendChild( links[0].cloneNode(true) ); // another link is added to the document
// the 'links' NodeList is automatically updated
// links.length === 3 now. If the NodeList is the return value of document.querySelectorAll, it is NOT live.
Going on this documentation, your current usage of the method indicates you do not have a live NodeList. Thus appending will never modify the original list. You will either need to modify your usage within the loop to mirror this usage of .cloneNode(true) or iterate manually.

How to make Javascript more specific?

I am new at JavaScript so I think my problem may be simple.
This works:
var convId = document.getElementById("wrapper");
convId.setAttribute("align","right");
But when I try to make it more specific:
var convId = document.getElementById("wrapper");
var convIdDl = convId.getElementsByTagName("dl");
convIdDl.setAttribute("align","right");
my definition list doesn't align to the right.
I have checked the HTML and CSS and everything is correct, but that shouldn't even matter
JavaScript overwrites them both.

The getElementsByTagName method returns a collection (to be more specific, a NodeList). You need to specify which element of that collection you want to use (just like you would when accessing an element in an array). Here I'm assuming you want the first:
convIdDl[0].setAttribute("align", "right");
As noted in the comments, you should definitely not be using the align attribute. CSS should be used in all cases.

The getElementsByTagName() function returns a collection of DOM elements, so you'll probably want to iterate through that and set the attribute on each element individually.
for(var i = 0; i < convIdDl.length; i++) {
convIdDl[i].setAttribute("align", "right");
}

Get list of all input objects using JavaScript, without accessing a form object

I need to get all the input objects and manipulate the onclick param.
The following does the job for <a> links. Looking for something like this for input tags.
for (var ls = document.links, numLinks = ls.length, i=0; i<numLinks; i++){
var link = unescape(ls[i].href);
link = link.replace(/\\'/ig,"#");
if(ls[i].href.indexOf("javascript:") == -1)
{
ls[i].href = "javascript:LoadExtern(\\""+link+"\\",\\"ControlPanelContent\\",true,true);";
}
}

(See update at end of answer.)
You can get a NodeList of all of the input elements via getElementsByTagName (DOM specification, MDC, MSDN), then simply loop through it:
var inputs, index;
inputs = document.getElementsByTagName('input');
for (index = 0; index < inputs.length; ++index) {
// deal with inputs[index] element.
}
There I've used it on the document, which will search the entire document. It also exists on individual elements (DOM specification), allowing you to search only their descendants rather than the whole document, e.g.:
var container, inputs, index;
// Get the container element
container = document.getElementById('container');
// Find its child `input` elements
inputs = container.getElementsByTagName('input');
for (index = 0; index < inputs.length; ++index) {
// deal with inputs[index] element.
}
...but you've said you don't want to use the parent form, so the first example is more applicable to your question (the second is just there for completeness, in case someone else finding this answer needs to know).
Update: getElementsByTagName is an absolutely fine way to do the above, but what if you want to do something slightly more complicated, like just finding all of the checkboxes instead of all of the input elements?
That's where the useful querySelectorAll comes in: It lets us get a list of elements that match any CSS selector we want. So for our checkboxes example:
var checkboxes = document.querySelectorAll("input[type=checkbox]");
You can also use it at the element level. For instance, if we have a div element in our element variable, we can find all of the spans with the class foo that are inside that div like this:
var fooSpans = element.querySelectorAll("span.foo");
querySelectorAll and its cousin querySelector (which just finds the first matching element instead of giving you a list) are supported by all modern browsers, and also IE8.

querySelectorAll returns a NodeList which has its own forEach method:
document.querySelectorAll('input').forEach( input => {
// ...
});
getElementsByTagName now returns an HTMLCollection instead of a NodeList. So you would first need to convert it to an array to have access to methods like map and forEach:
Array.from(document.getElementsByTagName('input')).forEach( input => {
// ...
});

var inputs = document.getElementsByTagName('input');
for (var i = 0; i < inputs.length; ++i) {
// ...
}