Match group only if it's different than other matching group

Question

I want to match every substring that begins with w and ends d with regex.

For example for input worldworld it should return ('worldworld', 'world', 'world'). (note: there are two world but they are different because they are at different position in the string)

For this purpose I ended with this program with following regex:

import re

s = '''worldworld'''

for g in re.finditer(r'(?=(w.*d))(?=(w.*?d))', s):
    print(g.start(1), g.end(1), g[1])
    print(g.start(2), g.end(2), g[2])
    print('-' * 40)

This prints:

0 10 worldworld
0 5 world
----------------------------------------
5 10 world
5 10 world
----------------------------------------

It finds all substrings, but some are duplicates also (notice the starting and ending position of the group).

I can filter the groups afterwards with group's starting and ending position, but I'm wondering if it can be done with change to my regex, to only return unique groups.

Can I change this regex to only match group that is different from other? If yes how? I'm open to suggestions how to solve this problem.

Answer 1

I don't believe this can be done with a single regexp. But it's straightforward with a nested loop:

import re
test = "wddddd"
# need to compile the tail regexp to get a version of
# `finditer` that allows specifying a start index
tailre = re.compile("(d)")
for wg in re.finditer("(w)", test):
    start = wg.start(1)
    for dg in tailre.finditer(test, wg.end(1)):
        end = dg.end(1)
        print(test[start : end], "at", (start, end))

That displays:

wd at (0, 2)
wdd at (0, 3)
wddd at (0, 4)
wdddd at (0, 5)
wddddd at (0, 6)

With

test = "worldworldworld"

instead:

world at (0, 5)
worldworld at (0, 10)
worldworldworld at (0, 15)
world at (5, 10)
worldworld at (5, 15)
world at (10, 15)

Answer 2

One option would be, with the lazy second group, to positive lookahead for .*d (greedy) afterwards to ensure that if the lazy second group matches, it's not the same as the greedy first group:

(?=(w.*d))(?:(?=(w.*?d)(?=.*d)))?

https://regex101.com/r/UI9ds7/2