How to match text in a non-greedy manner with regex for a specific recurrence

Question

With the regex pattern (Python):

(?<=start<).*?(?=>end)

I'd like to match/select only the most internal text of the following string:

start< obj1 obj2 start< obj3 >end

that is:

obj3

I use Pythex as online regex tester for my code (link). Pythex returns

obj1 obj2 start< obj3

instead of

obj3

Do you know a way to force the match of the most internal text? Maybe with some extra python code (if it is impossible with regex alone)?

Thanks

UPDATE 01 Sorry, I've tested your solutions (all) with different kinds of strings but I can't obtain what I want that is: match all between "start<" and ">end" but excluding strings containing "start<" and other characters before "start<".

For example if I have the string

start< obj1 >end start< obj2 >end start< obj3 start< obj4 >end

where "obj4" is equal to "<" (for example), no method/pattern proposed works because no method can match "<" at the end of the string. For the string above I'd like to obtain the following matched text (findall):

obj1
obj2
obj3
<

regardless of what "obj4" is (so the method I'm searching should work in general also if obj# is equal to "<").

Can you suggest some other solutions?

Thank you

could you simply add a space character before to make it `(?<= start<).*?(?=end)`? — SRT HellKitty, Oct 16 '19 at 15:07
Hi, thank you for reply, good answer for the original string, but doesn't work if we have `start< obj1 obj2 start< start< obj3 >end` (three `start<` instead of two). — Matteo VR, Oct 17 '19 at 07:39

score 0 · Answer 1 · answered Oct 16 '19 at 15:09

0

Use the following approach with improved regex pattern:

import re

s = 'start< obj1 obj2 start< obj3 >end'
m = re.search(r'(?<=start<)[^<]*?(?=>?end)', s)
res = m.group().strip() if m else m
print(res)    # obj3

answered Oct 16 '19 at 15:09

RomanPerekhrest

88,541
4
65
105

***On side note:-*** IMO `[^<]*?` no need of `?` as `[^<]*` this is already lazy by nature – Code Maniac Oct 16 '19 at 16:49
Hi, it works, thanks. Ok also the @CodeManiac version. I will try to play a little to understand this regex. – Matteo VR Oct 17 '19 at 07:51
@RomanPerekhrest How can I match also "<" if it is between "start<" and ">end"? – Matteo VR Oct 19 '19 at 11:47

score 0 · Answer 2 · answered Oct 16 '19 at 15:18

0

What about this?

r'start.*<(.*)>.*end'

In this case, the first * is so greedy that matches everything up to the last <. This gives you the inner text (accessible by group(1)).

answered Oct 16 '19 at 15:18

Patrizio G

362
3
13

Hi, good also this solution for the original string. I played a little and I've see that if we have the string `start< obj1 obj2 start< obj3 >end start< obj4 >end`, this regex matches only `obj4` and not also `obj3`, which is at the same depth level as `obj4`. – Matteo VR Oct 17 '19 at 08:00

score 0 · Answer 3 · answered Oct 16 '19 at 15:41

0

This can be done without lookahed or lookbehind, too:

s= "start< obj1 obj2 start< obj3 >end"
m=re.search(r"start<\s*([^<]*?)\s*>end",s)
>>> m[1]
>>> 'obj3'

answered Oct 16 '19 at 15:41

kantal

2,331
2
8
15

Hi, it works also for string like `start< obj1 >end >end >end obj2 start< start< obj3 >end start< start< obj4 >end`. I try to test this regex with other types of string similar and more complex to the original one. Thank you. – Matteo VR Oct 17 '19 at 08:03
@MatteoVR You mean using re.findall(), right? – kantal Oct 17 '19 at 10:41
Hi, yes I use re.findall() in my script. – Matteo VR Oct 18 '19 at 15:20
How can I match also "<" if it is between "start<" and ">end"? – Matteo VR Oct 19 '19 at 11:47

How to match text in a non-greedy manner with regex for a specific recurrence

3 Answers3