How to properly regex nested strings

Question

My program is trying to match strings surrounded by curly braces recursively - i.e when given {example1 {example2}} {example3} {example4} it will match example1 {example2}, example3 and example4. I tried using \{.+\}, but that will match example1 {example2}} {example3} {example4. \{.+?\} matches example1 {example2, example3 and example4. Is there an easy way to do what I want with a regex, or alternatively simply in Java? I am not very well-versed on regular expressions, as you can probably tell.

It is certainly possible, but it will be an absolute pain to code, since regexes are not meant for this. Just traverse the string and keep track of the number of `{` and `}`. — Keppil, Dec 08 '14 at 17:55
@AvinashRaj this isn't a duplicate, as I'm trying to extract the strings rather than ensure that brackets are balanced. — condorcraft110 II, Feb 09 '15 at 20:12
@Keppil, indeed, that's what I ended up doing - it was surprisingly little code — condorcraft110 II, Feb 09 '15 at 20:13

score 3 · Accepted Answer · edited Dec 08 '14 at 18:13

Regexs aren't good parsing things that require state. You could, probably, come up with something that works in some limited scenarios, but not a good regex that could handle a generic case.

You are better off just parsing your string explicitly. Basically, scan the characters one-by-one, and:

if the char is { increase the "nesting counter". If nesting counter==1, start a new group, and skip to next char;
if the char is }, decrease the counter. If nesting_counter == 0, end the current group, and skip to next char;
if nesting_counter>0 append the current char to the group and continue;

How to properly regex nested strings

1 Answers1