Regular expression to extract text between \()

Question

I have a string on the following format:

abc\(def\(ghj)klm)nop\(qrs)tuvw\(xyz\(abc\(def)ghi)jkl)mno)pqr)\(stu)

My current regex is :

(\\\((?:\[??[^\"\n]*?\)))

and it is give me the output in the image, how can I update this regex to achieve the expectation

Regex aren't suited to parse tree-hierarchy structures. Although this *may* be possible, you'll end up with a complicated RegEx hard to maintain and easy to break — Cid, Mar 11 '21 at 07:29
Sounds way easier to match the number of `\(` with the number of `)` using a *good old loop* — Cid, Mar 11 '21 at 07:30
I second the loop approach. What language do you use? If regex, how many levels deep? If, say max 3 levels, you could build a longer regex that first matches for 3 levels, then 2 levels, then 1 level. — Peter Thoeny, Mar 11 '21 at 08:05

score 1 · Answer 1 · answered Mar 11 '21 at 08:17

1

Recursion can helps you with that:

\\\(([^()]|(?R))*\)

Example here

answered Mar 11 '21 at 08:17

Vladimir Trifonov

39
6

Thank you!! I use Swift language and it seems not support Recursion with regex. – Hassan Taleb Mar 11 '21 at 09:20

Peter Thoeny · Answer 2 · 2021-03-11T19:12:11.347

1

There are regex dialects that support recursions, namely Perl 5.10, PCRE 4.0, Ruby 2.0. You can match your string with a regex engine that supports recursion as @Vladimir Trifonov pointed out in his answer.

If your regex engine does not support recursion you can match without recursion by building a longer regex that supports up to a max level of nesting, here using your input string as an example:

max 1 level: /(\\\([^\)]*\))/g
- match: ["\(def\(ghj)", "\(qrs)", "\(xyz\(abc\(def)", "\(stu)"]
max 2 levels: /(\\\((?:[^\\\)]*\\\([^\)]*\))?[^\)]*\))/g
- match: ["\(def\(ghj)klm)", "\(qrs)", "\(xyz\(abc\(def)ghi)", "\(stu)"]
max 3 levels: /(\\\((?:[^\\\)]*\\\((?:[^\\\)]*\\\([^\)]*\))?[^\)]*\))?[^\)]*\))/g
- match: ["\(def\(ghj)klm)", "\(qrs)", "\(xyz\(abc\(def)ghi)jkl)", "\(stu)"]
etc.

In other words, you can increase the max level by nesting (?:[^\\\)]*\\\( ... [^\)]*\))? patterns; they are made optional with ?.

EDIT: In case you need to support unlimited nesting and your regex engine does not support recursion, you can use a series of regexes. See JavaScript example at https://stackoverflow.com/a/66558399/7475450 and Perl example at https://twiki.org/cgi-bin/view/Blog/BlogEntry201109x3

edited Mar 11 '21 at 19:12

answered Mar 11 '21 at 08:30

Peter Thoeny

7,379
1
10
20

Thank you!! I use Swift language and it seems not support Recursion with regex. the problem is that there is no max number for level. – Hassan Taleb Mar 11 '21 at 09:20
@HassanTaleb: YW. In this case how about rinse & repeating that nesting to, say, 10 levels, and document that as a limitation? – Peter Thoeny Mar 11 '21 at 09:30
I use this regex for highlighting syntax in an editor that I created. So since user can write any syntax I can't limit levels. – Hassan Taleb Mar 11 '21 at 10:13
@HassanTaleb: You could expand that regex to a level that is unlikely to be reached, like 20. Alternatively, see **EDIT** in my answer. – Peter Thoeny Mar 11 '21 at 19:13
@HassanTaleb: Any questions? Did you find a solution? – Peter Thoeny Mar 14 '21 at 01:50

Regular expression to extract text between \()

2 Answers2