正则表达式在Javascript替代方法中落后

Question

I'm trying to use the following regex in JS:

(?<=@[AZ|az]+,)\\s|(?<=@[AZ|az]+,\\s[AZ|az]+)\\s(?=\\[[AZ|az]+\\])

which translates to:

match all spaces which are preceded by :

• @
• followed by any number of characters in the range AZ or az
• followed by a comma

OR

match all spaces which are preceded by:

• @

• followed by any number of characters in the range AZ or az

• followed by a comma
• followed by a space
• followed by any number of characters in the range AZ or az

AND are succeeded by:

• [
• followed by any number of characters in the range AZ or az
• ]

However, JS doesn't support lookbehind. Is there any alternative for supporting the above regex in JS or any npm library I can use instead?

So if we have a sentence like
Hi my name is @John, Doe [Example] and I am happy to be here that should become
Hi my name is @John,Doe[Example] and I am happy to be here .
Also, if we have something like
Hi my name is @John, Smith Doe [Example] , that should become
Hi my name is @John,SmithDoe[Example] .

Answer 1

I've updated my answer on new input

 console.clear(); var inputEl = document.querySelector('#input') var outputEl = document.querySelector('#output') function rep (e) { var input = e.target.value; var reg = /@([az]+?\\s*?)+,(\\s+[az]+)+(\\s\\[[az]+\\])?/gim matches = input.match(reg); var output = input; if (matches) { replaceMap = new Map() for (var i = 0; i < matches.length; i++) { var m = matches[i] .replace(/\\[/, '\\\\[') .replace(/\\]/, '\\\\]') replaceMap.set(m, matches[i].replace(/\\s+/gm, '')) } for (var [s,r] of replaceMap) { output = output.replace(new RegExp(s, 'gm'), r) } } outputEl.textContent = output } inputEl.addEventListener('input', rep) inputEl.dispatchEvent(new Event('input')) 

 textarea { width: 100%; min-height: 100px; } 

 <h3>Input</h3> <textarea id="input">@Lopez de la Cerda, Antonio Gabriel Hugo David [Author]. I'm the father of @Marquez, Maria</textarea> <h3>Output (initially empty)</h3> <p id="output"></p> <h3>Expected result (on initial input)</h3> <p>@LopezdelaCerda,AntonioGabrielHugoDavid[Author]. I'm the father of @Marquez,Maria</p> 

Backup of old answer content (for historical reasons)

It works at least in Chrome with this regex:

/(?<=@[a-z]+,)\s+(?![a-z]+\s+\[[a-z]+\])|(?<=(@[a-z]+,\s[a-z]+))\s+(?=\[[a-z]+\])/gmi

See: https://regex101.com/r/elTkRe/4

But you can't use it in PCRE because it is not allowed to have quantifiers in lookbehinds. They must be of fixed width. See the errors to the right here: https://regex101.com/r/ZC3XmX/2

Solution without look behinds and look aheads

 console.clear(); var reg = /(@[A-Za-z]+,\\s[A-Za-z]+)(\\s+)(\\[[A-Za-z]+\\])|(@[AZ|az]+,)(\\s+)/gm var probes = [ '@gotAMatch, <<<', '@LongerWithMatch, <<<', '@MatchHereAsWell, <<<', '@Yup, <<<<', '@noMatchInThisLine,<<<<<', '@match, match [match]<<<<<<<', '@ noMatchInThisLine, <<<<' ] for (var i in probes) { console.log(probes[i].replace(reg, '$1$3$4')) } 

 .as-console-wrapper { max-height: 100% !important; top: 0; } 

Answer 2

What you need to do is converting lookbehinds to capturing groups in order to include them in replacement string (note that case-insensitive flag ( i ) is set):

(@[a-z]+,)([\t ]*([a-z]+)[\t ]*(?=\[[a-z]+\])|[\t ]+)

Replace with $1$3 if you want to remove those spaces.

See live demo here

Answer 3

Just update your Node.js version. Lookbehind assertions are part of ECMAScript 2018 and are already implemented in Chromium and Node.js. According to http://kangax.github.io/compat-table/es2016plus/ , Chromium 70 and Node.js 8.10 have this feature.

I just tested it in my browser and in Node.js (v8.11) and can confirm that:

node -e "console.log('nothing@xyz, bla'.match(/(?<=@[A-Za-z]+,)\s+/))"

If you can't update you have to use other strategies like capture and replace which should not be a big issue with a positive lookbehind (negatives are harder):

const hit = 'nothing@xyz, bla'.match(/(@[A-Za-z]+,)\s+/)
hit[0].replace(hit[1])

If nothing else works, take a look at this project which tries to implement Lookbehind (I haven't tested it): http://blog.stevenlevithan.com/archives/javascript-regex-lookbehind