Regex: Everything until last occurrence of pattern

Question

Given the following strings:

'/a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f/g'
'/a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f/'
'/a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f'
'/a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f/g/xxx/xxx'

I'm looking for a regex that will match

/a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f

for all of the above strings

That is, match everything until the last occurrence of this pattern:

{forward slash or start of string} then {xxx} then {forward slash} then {any group of chars (except forward slash) that is not equal to xxx } then {forward slash or end of string}

I do not know how to translate the above pattern into a regex.

The .*\/ pattern I tried does not solve the problem.

@WiktorStribiżew Thanks for your always good regexs. I think this isn't a dupplicate of what you linked, because that answers include the /. — Jorge.V, Jul 06 '18 at 09:09
@WiktorStribiżew nevermind, OP edited question and fully changed it. — Jorge.V, Jul 06 '18 at 09:10
@SproutCoder What is the code/pattern that does not work for you? Your current title and *match everything until the last occurrence of this pattern* requirement is in line with [this post](https://stackoverflow.com/questions/24465512/regular-expression-to-match-everything-until-the-last-occurrence-of). Please update the question if that solution does not work for you. — Wiktor Stribiżew, Jul 06 '18 at 09:11
@SproutCoder I don't understand your edit. Why does /a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f match entirely and not /a/xxx/b/c/xxx/xxx/d/e/xxx/xxx since the last /whatever shouldn't be included? And where is the g for the last supposed good match? Shouldn't it just remove the /xxx.. part? In fact shouldn't it remove only the last /xxx? — Jorge.V, Jul 06 '18 at 09:14
So, https://regex101.com/r/UYvZDh/6 works as expected,right? — Wiktor Stribiżew, Jul 06 '18 at 11:36
yes it looks like this is the correct answer, I can't think of any other edge cases, it looks like this is following the exact pattern I mentioned — Kawd, Jul 06 '18 at 11:39
Ok, note that `{anything but xxx}` is not clear in the question, it can be understood as `(?:(?!xxx).)*` to match any char, 0+ times, that does not start `xxx` sequence. And what you need is `{any subpart not equal to xxx}`. — Wiktor Stribiżew, Jul 06 '18 at 11:41

score 1 · Accepted Answer · answered Jul 06 '18 at 11:41

You may use

/(?:.*\/)?xxx\/(?!xxx(?![^\/]))[^/]*(?=$|\/)/

See the regex demo.

Details

(?:.*\/)? - an optional group matching
- .* - any 0+ chars other than line break chars, as many as possible
- \/ - a / char
xxx\/ - a xxx/ substring
(?!xxx(?![^\/])) - the text starting at the next location in string cannot be equal to xxx/ or xxx at the end of the string
[^/]* - any 0+ chars other than / as many as possible ([^/]*)
(?=$|\/) - there must be end of string or / immediately to the right of the current position.

JS demo:

var strs = ['/a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f/g','/a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f/','/a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f','/a/xxx/b/c/xxx/xxx/d/e/xxx/xxx/f/g/xxx/xxx','/xxx/f/g','xxx/f/g','a/xxx/f/g','/a/xxx/f/g','/a/xxx/b/c/xxx/xxx/d/e/xxx/gggxxxd/f/gxxxf/'];
var rx = /(?:.*\/)?xxx\/(?!xxx(?![^\/]))[^/]*(?=$|\/)/;
for (var s of strs) {
  console.log(s, "=>", s.match(rx)[0]);
}

Regex: Everything until last occurrence of pattern

1 Answers1