How make a conditional split in Javascript?

Question

I need a split conditional to split by ;

Except IF the ; is after a &lt ou &gt. (< or > are converted in < or >)

Example:

Input:

A &gt; , &lt; B , &lt; C &gt; ; 100, 119, 150

Output:

A &gt;, &lt; B, &lt; C &gt;
100,119,150

I can make it:

var myString =  "A &gt; , &lt; B , &lt; C &gt; ; 100, 119, 150"
myString.replace("&gt;", "!!!before").replace("&lt;", "!!!after");    
myString.split(";");
myString.replace("!!!before", "&gt;").replace("!!!after", "&lt;");

But this solution is a big workaround to my case, I'm searching a solution using regex or indexof to improve it.

I don't suppose you can just resolve the HTML entities *before* the split? — Touffy, Apr 15 '15 at 18:53
I cannot resolve before, because this input is used to a API. I can make two replaces (like replace !! instead <, make my split by ; and after replace again), but I'm searching a solution using regex, for example. — Helio Bentzen, Apr 15 '15 at 18:55
Your question is receiving negative votes because you've shown no effort that you've made. — Etheryte, Apr 15 '15 at 18:55
@HelioBentzen You get the negative votes because your question doesn't show any research effort. Show us what you have tried so far, explain why it didn't work, etc. — Gergo Erdosi, Apr 15 '15 at 18:56
too bad [there's no lookbehind](http://stackoverflow.com/questions/641407/javascript-negative-lookbehind-equivalent) in JavaScript regexes — Touffy, Apr 15 '15 at 19:12

Etheryte · Accepted Answer · 2015-04-15T19:05:37.537

1

One easy solution would be to use placeholders that you shouldn't otherwise expect in the given context. I've used unicode zero-width characters in the below example:

var arr = "A &gt; , lgt; B , lgt; C &gt; ; 100, 119, 150"
    .replace(/&gt;/g, "\u200B")
    .replace(/&lt;/g, "\u200C")
    .split(";");
arr.forEach(function(el, i) {
    arr[i] = el.replace(/\u200B/g, "&gt;").replace(/\u200C/g, "&lt;");
});
console.log(arr); //outputs ["A &gt; , lgt", " B , lgt", " C &gt; ", " 100, 119, 150"]

Addressing the update you added to your question: Despite regex occasionally looking shorter, it usually offers worse performance by far, see for example this article on Coding Horror.

edited Apr 15 '15 at 19:05

answered Apr 15 '15 at 18:55

Etheryte

24,589
11
71
116

no need to use placeholders, the actual < and > characters would do fine. – Touffy Apr 15 '15 at 18:56
@Touffy Not if you have the chars in the input string. Surely, you wouldn't expect this from something that's escaped, but the question didn't specify, there might be other operations going on before reaching this state as well. Regardless of the placeholders, this shows an easy approach of how to solve the problem. – Etheryte Apr 15 '15 at 18:58
@Touffy, I really cannot use < and > – Helio Bentzen Apr 15 '15 at 19:04
@HelioBentzen See the additional comment I appended to my answer regarding using regex. – Etheryte Apr 15 '15 at 19:05
1

@Nit, so, it's a great argument to avoid regex. Thanks! – Helio Bentzen Apr 15 '15 at 19:09

How make a conditional split in Javascript?

1 Answers1