I have the following situation, I'm searching in a HTML string for the attributes.

I have the following regex which works alright, but I want to get just unique results, of course I can apply some filter to the results array but I think this is achievable with pure regex.

So in this situation class is returned twice, but I only want 1 time:

['class', 'data-text'] not ['class', 'data-text', 'class']

const html = `<div class="foo">
    <span data-text="Some string" class="bar"></span>
</div>`

console.log(html.match(/[\w-:]+(?=\s*=\s*".*?")/g))

I have the following situation, I'm searching in a HTML string for the attributes.

I have the following regex which works alright, but I want to get just unique results, of course I can apply some filter to the results array but I think this is achievable with pure regex.

https://regex101./r/UqCuJS/1

So in this situation class is returned twice, but I only want 1 time:

['class', 'data-text'] not ['class', 'data-text', 'class']

const html = `<div class="foo">
    <span data-text="Some string" class="bar"></span>
</div>`

console.log(html.match(/[\w-:]+(?=\s*=\s*".*?")/g))

http://jsbin./bekibanisa/edit?js,console

• javascript
• regex

• Can you include full javascript tried at Question? Are you using .split() or .match() to get array of matches? – guest271314 Commented Apr 30, 2017 at 15:15
• jsbin./bekibanisa/edit?js,console check this out, .match – Anderson Commented Apr 30, 2017 at 15:16
• 2 See jsbin./yofozabilu/1/edit?js,console – Wiktor Stribiżew Commented Apr 30, 2017 at 15:19
• @WiktorStribiżew your answer seems right, can you post it with an explanation of what is happening? – Anderson Commented Apr 30, 2017 at 15:21
• 1 Yes, the greedy patterns with a construct matching any char, any number of times, will cause much backtracking. Use /[\w-:]+(?=\s*=\s*"[^"]*")/g and use .filter(). – Wiktor Stribiżew Commented Apr 30, 2017 at 15:49

2 Answers 2

You can pass result of .match() to Set, which does not allow duplicate values. If necessary convert Set instance back to Array.

const html = `<div class="foo">
	<span data-text="Some string" class="bar"></span>
</div>`
// or use existing `RegExp`
console.log([...new Set(html.match(/([\w-]+)(?=[=]")/g))])

Try removing the '/g' global modifier

console.log(html.match(/[\w-:]+(?=\s*=\s*".*?")/))