Match pattern except under one condition Regex

Question

I'm trying to match a patterned with regex except when the pattern is escaped.

Test text:

This is AT\&T&reg; is really cool Regex

You can see with my \& I'm manually escaping. And therefore, do not want the regex to match.

Regex:

const str = 'This is AT\&T&reg; is really cool Regex'

str.replace(/\&(.*?)\;/g, '<sup>&$1;</sup>');

Expected output

This is AT&T<sup>&reg;</sup> is really cool Regex

Hard to explain I guess but when the start of this regex looks for a & and ends with a ; however, if & is preceded with at \ like \& than do not match and look for the next \&(.*?)\;

Your `str` does not contain a literal ``\``. You need to double ``\`` inside a string literal to define a literal backslash. — Wiktor Stribiżew, Dec 30 '17 at 18:57
I want \& to never match and go on to lookup the next &(.*?)\; — Armeen Moon, Dec 30 '17 at 19:00
A better example, I guess, would be `var s = "This is AT\\®® is really cool Regex";`. Your current problem can be solved with `/&\w+;/g` regex. — Wiktor Stribiżew, Dec 30 '17 at 19:03
This is not a question, this is "write code for me". What's worse, you didn't say what language you're using and the task is underspecified (e.g. you haven't answered transistor09's question yet). — melpomene, Dec 30 '17 at 19:07

Anton · Accepted Answer · 2017-12-30T19:34:13.923

1

You can use negative lookbehind This regex works fine with the example /(?<!\\)\&(.*?)\;/g

Edit 1

To workaround in JS you can use [^\\] that will match everything except backslash. The overall regex /[^\\]\&(.*?)\;/g It works for your example.

edited Dec 30 '17 at 19:34

answered Dec 30 '17 at 19:02

Anton

442
4
19

Webstorm IDE hates it but I love you! Thanks – Armeen Moon Dec 30 '17 at 19:07
@MatthewHarwood JavaScript doesn't have look-behind. – melpomene Dec 30 '17 at 19:08
JavaScript [doesn't seems to support](https://developer.mozilla.org/en-US/docs/Web/JavaScript/Guide/Regular_Expressions) lookbehind assertions so it should be tested. – Flying Dec 30 '17 at 19:08
Sorry, I didn't notice that it is JS. There is workaround for JS https://stackoverflow.com/a/11347100/1079966 (also check out other answers in this thread) – Anton Dec 30 '17 at 19:09
@Anton The JS tag was only added a minute ago. – melpomene Dec 30 '17 at 19:10
`str.replace` clearly points that it is JS snippet, but tag was missed, true – Flying Dec 30 '17 at 19:17

score 0 · Answer 2 · answered Dec 30 '17 at 19:16

Since JavaScript have no support for lookbehind assertions - it is possible to add some custom substitution logic to achieve desired results. I've updated test string with examples of different kinds of html entities for test purposes:

const str = '&T;his is AT\\&T&reg; is & really &12345; &xAB05; \\&cool; Regex'

console.log(str.replace(/&([a-z]+|[0-9]{1,5}|x[0-9a-f]{1,4});/ig, function (m0, m1, index, str) {
    return (str.substr(index - 1, 1) !== '\\') ? '<sup>' + m0 + '</sup>' : m0;
}));

Match pattern except under one condition Regex

2 Answers2

Edit 1