Develop Reference

jQuery closest (inside and outside DOM tree )

Given an element and any selector, I need to find the closest element which matches it, not matter if it's inside the element or outside of it.
Currently jQuery doesn't provide such traversing functionality, but there is a need. Here is the scenario:
A list of many items where the <button> element reside inside <a>
<ul>
<li>
<a>
<button>click me</button>
<img src="..." />
</a>
</li>
<li>
<a>
<button>click me</button>
<img src="..." />
</a>
</li>
...
</ul>
Or the <button> element might reside outside of the <a> element
<ul>
<li>
<a>
<img src="..." />
</a>
<button>click me</button>
</li>
<li>
<a>
<img src="..." />
</a>
<button>click me</button>
</li>
...
</ul>
The very very basic code would look like this:
$('a').closest1('button'); // where `closest1` is a new custom function
// or
$('a').select('> button') // where `select` can parse any selector relative to the object, so it would also know this:
$('a').select('~ button') // where the button is a sibling to the element
the known element is <a> and anything else can change. I want to locate the nearest <button> element for a given <a> element, no matter if that button is inside or outside of <a>'s DOM tree.
It would be very logical that native jQuery function "closest" would do as the name suggests and find the closest, but it only searches upwards as you all know. (it should have been named differently IMO).
Does anyone know any custom traversing function which does the above?
Thanks. (i'm asking you people because someone must have written this for sure but I was unlucky to find a lead on the internet)

Here is another attempt using the idea I mentioned in comment:
$(this).parents(':has(button):first').find('button').css({
"border": '3px solid red'
});
JSFiddle: http://jsfiddle.net/TrueBlueAussie/z3vwk1ko/40/
It basically looks for the first ancestor that contains both the elements (clicked and target), then finds the target.
Performance:
With regard to speed, this is used at human interaction speeds, i.e. a few times per second maximum, so being a "slow selector" is irrelevant if it solves the problem, in a reasonably obvious way, with minimal code. You would have to click 100s of times per second to notice any different compared to a fast selector :)

None of the built-in selectors allow searching up and down the tree. I did create a custom findThis extension that allows you to do things like $elementClicked.('li:has(this) button') which would allow you to do something similar.
// Add findThis method to jQuery (with a custom :this check)
jQuery.fn.findThis = function (selector) {
// If we have a :this selector
if (selector.indexOf(':this') > 0) {
var ret = $();
for (var i = 0; i < this.length; i++) {
var el = this[i];
var id = el.id;
// If not id already, put in a temp (unique) id
el.id = 'id' + new Date().getTime();
var selector2 = selector.replace(':this', '#' + el.id);
ret = ret.add(jQuery(selector2, document));
// restore any original id
el.id = id;
}
ret.selector = selector;
return ret;
}
// do a normal find instead
return this.find(selector);
}
// Test case
$(function () {
$('a').click(function () {
$(this).findThis('li:has(:this) button').css({
"border": '3px solid red'
});
});
});
JSFiddle: http://jsfiddle.net/TrueBlueAussie/z3vwk1ko/38/
note: Click the images/links to test.

A while ago I wanted to do the same for a completely DOM-based text editor, and needed to find the previous (ARR LEFT) and next (ARR RIGHT) text nodes, both up and down the tree. Based on this code I have made an adaptation suiting this question. Be warned, it's quite performance-heavy, but it is adapted to any scenario. There are two functions findPrevElementNode and findNextElementNode which both return an object with properties:
match - returns the closest matching node for the search or FALSE if none is found
iterations - returns the number of iterations done to find the node. This allows you to check whether the previous node is closer than the next or vice-versa.
The parameters are as follows:
//#param {HTMLElement} referenceNode - The node from which to start the search
//#param {function} truthTest - A function that returns true for the given element
//#param {HTMLElement} [limitNode=document.body] - The limit up to which to search to
var domUtils = {
findPrevElementNode: function(referenceNode, truthTest, limitNode) {
var element = 1,
iterations = 0,
limit = limitNode || document.body,
node = referenceNode;
while (!truthTest(node) && node !== limit) {
if (node.previousSibling) {
node = node.previousSibling;
iterations++;
if (node.lastChild) {
while (node.lastChild) {
node = node.lastChild;
iterations++;
}
}
} else {
if (node.parentNode) {
node = node.parentNode;
iterations++;
} else {
return false;
}
}
}
return {match: node === limit ? false : node, iterations: iterations};
},
findNextElementNode: function(referenceNode, truthTest, limitNode) {
var element = 1,
iterations = 0,
limit = limitNode || document.body,
node = referenceNode;
while (!truthTest(node) && node !== limit) {
if (node.nextSibling) {
node = node.nextSibling;
iterations++;
if (node.firstChild) {
while (node.firstChild) {
node = node.firstChild;
iterations++;
}
}
} else {
if (node.parentNode) {
node = node.parentNode;
iterations++;
} else {
return false;
}
}
}
return {match: node === limit ? false : node, iterations: iterations};
}
};
In your case, you could do:
var a = domUtils.findNextElementNode(
document.getElementsByTagName('a')[0], // known element
function(node) { return (node.nodeName === 'BUTTON'); }
);
var b = domUtils.findPrevElementNode(
document.getElementsByTagName('a')[0], // known element
function(node) { return (node.nodeName === 'BUTTON'); }
);
var result = a.match ? (b.match ? (a.iterations < b.iterations ? a.match :
(a.iterations === b.iterations ? fnToHandleEqualDistance() : b.match)) : a.match) :
(b.match ? b.match : false);
See it in action.

DEMO PAGE / GIST
I have solved working by logic, so I would first look inside the elements, then their siblings, and last, if there are still unfound items, I would do a recursive search on the parents.
JS CODE:
jQuery.fn.findClosest = function (selector) {
// If we have a :this selector
var output = $(),
down = this.find(selector),
siblings,
recSearch,
foundCount = 0;
if(down.length) {
output = output.add(down);
foundCount += down.length;
}
// if all elements were found, return at this point
if( foundCount == this.length )
return output;
siblings = this.siblings(selector);
if( siblings.length) {
output = output.add(siblings);
foundCount += siblings.length;
}
// this is the expensive search path if there are still unfound elements
if(foundCount < this.length){
recSearch = rec(this.parent().parent());
if( recSearch )
output = output.add(recSearch);
}
function rec(elm){
var result = elm.find(selector);
if( result.length )
return result;
else
rec(elm.parent());
}
return output;
};
// Test case
var buttons = $('a').findClosest('button');
console.log(buttons);
buttons.click(function(){
this.style.outline = "1px solid red";
})

I think using sibling selector (~) or child selector (>) will solve your purpose(What ever your case is!!).

How do I search nested divs using regex for class names

I want to search for nested divs like these in my DOM
<div class="two-columns some-other-class">
<div class="two-columns some-other-class">
</div>
</div>
I tried to search nested divs like this
$("div[class$='columns'] div[class$='columns']")
but it's not working :(

Use the Attribute Contains Selector:
$("div[class*='columns'] div[class*='columns']")
Edit:
If you want that exact functionality you might want to extend the jQuery selector engine:
$.extend($.expr[':'], {
classEndsWith: function(e, i, meta) {
var found = false
var classes = e.className.toLowerCase().split(' ');
var suffix = meta[3].toLowerCase();
$.each(classes, function(i, c) {
// Check if className ends with value in suffix
if (c.indexOf(suffix, c.length - suffix.length) !== -1) {
found = true;
return false;
}
});
return found;
}
});
var element = $('div:classEndsWith(columns) div:classEndsWith(columns)');
See JSFiddle.

$("div[class$='columns'] div[class$='columns']")
Is working. Check the fiddle

Convert URL in paragraphs to links using jQuery or Javascript

I am new to javascript but understand jQuery. I am trying to use this code to convert www. and http in p tags to working links.
Here is the code I am using, the problem is that I do not fully understand how the code works, could anybody please explain?
<script>
var re = /(http:\/\/[^ ]+)/g;
function createLinks(els) {
$(els).contents().each(function () {
if (this.nodeType === 1 && this.nodeName !== 'script') {
createLinks(this);
} else if (this.nodeType === 3 && this.data.match(re)) {
var markup = this.data.replace(re, '$1');
$(this).replaceWith(markup);
}
});
}
createLinks(document.body);
</script>

First, you set regular expression template for matching text which starts from "http://"
Second, you create recursive function which traverse whole html document.
nodeType == 1 means that current element is html tag (i.e. a, p, div etc)
nodeType == 2 means that element is Attribute
nodeType == 3 means that element is text node
So when you found html tag, you're searching inside it,
when you found text node, you are checking via regular expression, if this text starts from "http://", if so you change and replce this text to yourmatchedurl
in the end you call your function to start from body as a root

ok, here goes...
//create a regular expression to format the link
var re = /(http:\/\/[^ ]+)/g;
//this is the create links function which gets called below, "els" is the elements passed to the function (document.body)
function createLinks(els) {
//for each of the elements in the body
$(els).contents().each(function () {
//check if its an element type but not a script
if (this.nodeType === 1 && this.nodeName !== 'script') {
//call the create links function and send in this object
createLinks(this);
//if its not an element but is a text node and the format matches the regular expression
} else if (this.nodeType === 3 && this.data.match(re)) {
//create the markup
var markup = this.data.replace(re, '$1');
//finally, replace this link with the marked up link
$(this).replaceWith(markup);
}
});
}
//call the create links function
createLinks(document.body);
I hope the commented code helps you understand.

jQuery selector for an element that directly contains text?

I was able to get this partially working using the :contains selector, but my problem is if an element contains an element that contains the text it is still returned. For example:
$('div:contains("test")')
Will select both divs below:
<div>something else
<div>test</div>
</div>
fiddle: http://jsfiddle.net/TT7dR/
How can I select only divs that "directly" contain the text? Meaning that in the above example only the child div would be selected.
UPDATE:
Just to clarify, if I were searching for the text "something else" instead of "test" then I would like to only find the parent div.

$('div>:contains("test")') is not a general solution, it only works for your specific example. It still matches any element whose descendants contain the text test, as long as its parent is a div.
There is in fact currently no selector that will select only direct parents of text nodes containing your target text. To do it you would have to walk the DOM tree yourself checking each text node you find for the target text, or write a plugin to do the same. It'd be slow, but then not as slow as :contains already is (it's not a standard CSS selector so you don't get browser-native fast selector support).
Here's a plain DOM function you could use as a starting point. It might be improved to find text in adjacent (non-normalised) text nodes, or to hide it in a plugin/selector-extension.
function findElementsDirectlyContainingText(ancestor, text) {
var elements= [];
walk(ancestor);
return elements;
function walk(element) {
var n= element.childNodes.length;
for (var i= 0; i<n; i++) {
var child= element.childNodes[i];
if (child.nodeType===3 && child.data.indexOf(text)!==-1) {
elements.push(element);
break;
}
}
for (var i= 0; i<n; i++) {
var child= element.childNodes[i];
if (child.nodeType===1)
walk(child);
}
}
}

Just to complete the knowledge base. If you need to get all DOM elements within the body (not only DIVs) that contain specific text or characters you can use:
function getNodesThatContain(text) {
var textNodes = $(document).find(":not(iframe, script)")
.contents().filter(
function() {
return this.nodeType == 3
&& this.textContent.indexOf(text) > -1;
});
return textNodes.parent();
};
console.log(getNodesThatContain("test"));
Hope that helps.
jsfiddle: http://jsfiddle.net/85qEh/2/
Credits: DMoses

You might have to do an in-efficient query. Do not use this solution if someone finds a selector that manages to filter out child elements: http://viralpatel.net/blogs/2011/02/jquery-get-text-element-without-child-element.html
$("div:contains('test')")
.clone() //clone the element
.children() //select all the children
.remove() //remove all the children
.end() //again go back to selected element
.filter(":contains('test')")
edit: that snippet above is just to test the element, in implementation it would look more like this: http://jsfiddle.net/rkw79/TT7dR/6/
$("div:contains('test')").filter(function() {
return (
$(this).clone() //clone the element
.children() //select all the children
.remove() //remove all the children
.end() //again go back to selected element
.filter(":contains('test')").length > 0)
}).css('border', 'solid 1px black');

try adding the greater than:
$('div>:contains("test")')

Finds specific element, but not parents
var elementsContainingText = ($(':contains("' + text + '")', target)).filter(function() {
return $(this).contents().filter(function() {return this.nodeType === 3 && this.nodeValue.indexOf(text) !== -1; }).length > 0;
});

This seems to work for me:
$('div >:contains("test")');
http://jsfiddle.net/TT7dR/1/
This forces the matched :contains selector to be a direct child of the <div> element

Try the following:
$("div>div:contains(test):only-of-type")

Add more alternative:
if($(selector).text().trim().length) {
var thetext = $(selector).contents().filter(function(){
return this.nodeType === 3;
}).text().trim();
console.log(thetext);
}
It will select the text only and remove any element with tag!
Reference

You can simply select the element that doesn't have your element
$('div:contains("test"):not(:has(> div))')

less code to write (but with a little limitation):
let selector = $('div:contains("test")');
selector.not(selector.has('div:contains("test")'))
Just use the jQuery function (.has) because the css :has is experimental:
https://developer.mozilla.org/en-US/docs/Web/CSS/:has#Browser_compatibility
Limitation:
When you have a structure like this:
<div>
<div>test</div>
test
</div>
Then only the inner div - Element will be found by this solution. This is because there is still an Element - Child of the div that :contains the string "test".

How to select a part of string?

How to select a part of string?
My code (or example):
<div>some text</div>
$(function(){
$('div').each(function(){
$(this).text($(this).html().replace(/text/, '<span style="color: none">$1<\/span>'));
});
});
I tried this method, but in this case is selected all context too:
$(function(){
$('div:contains("text")').css('color','red');
});
I try to get like this:
<div><span style="color: red">text</span></div>

$('div').each(function () {
$(this).html(function (i, v) {
return v.replace(/foo/g, '<span style="color: red">$&<\/span>');
});
});

What are you actually trying to do? What you're doing at the moment is taking the HTML of each matching DIV, wrapping a span around the word "text" if it appears (literally the word "text") and then setting that as the text of the element (and so you'll see the HTML markup on the page).
If you really want to do something with the actual word "text", you probably meant to use html rather than text in your first function call:
$('div').each(function(){
$(this).html($(this).html().replace(/text/, '<span style="color: none">$1<\/span>'));
// ^-- here
}
But if you're trying to wrap a span around the text of the div, you can use wrap to do that:
$('div').wrap('<span style="color: none"/>');
Like this: http://jsbin.com/ucopo3 (in that example, I've used "color: blue" rather than "color: none", but you get the idea).

$(function(){
$('div:contains("text")').each(function() {
$(this).html($(this).html().replace(/(text)/g, '<span style="color:red;">\$1</span>'));
});
});
I've updated your fiddle: http://jsfiddle.net/nMzTw/15/

The general practice of interacting with the DOM as strings of HTML using innerHTML has many serious drawbacks:
Event handlers are removed or replaced
Opens the possibility of script inject attacks
Doesn't work in XHTML
It also encourages lazy thinking. In this particular instance, you're matching against the string "text" within the HTML with the assumption that any occurrence of the string must be within a text node. This is patently not a valid assumption: the string could appear in a title or alt attribute, for example.
Use DOM methods instead. This will get round all the problems. The following will use only DOM methods to surround every match for regex in every text node that is a descendant of a <div> element:
$(function() {
var regex = /text/;
function getTextNodes(node) {
if (node.nodeType == 3) {
return [node];
} else {
var textNodes = [];
for (var n = node.firstChild; n; n = n.nextSibling) {
textNodes = textNodes.concat(getTextNodes(n));
}
return textNodes;
}
}
$('div').each(function() {
$.each(getTextNodes(this), function() {
var textNode = this, parent = this.parentNode;
var result, span, matchedTextNode, matchLength;
while ( textNode && (result = regex.exec(textNode.data)) ) {
matchedTextNode = textNode.splitText(result.index);
matchLength = result[0].length;
textNode = (matchedTextNode.length > matchLength) ?
matchedTextNode.splitText(matchLength) : null;
span = document.createElement("span");
span.style.color = "red";
span.appendChild(matchedTextNode);
parent.insertBefore(span, textNode);
}
});
});
});