there will always be only one or zero occurances of '#' so there will only ever be two strings

spoiler flag, thx...

You may want to have some unit test before you refactor your code. Under a good protection, changing any implementation should be safe :)

Pro tip: don't waste your energy trying to up the quality of codewars posts

Bad practice because:

1. It finds all occurrences of '#' in string
2. Gets all possible substrings
3. Takes just one of them.

If it is faster, why not just give it a shot?

I implemeted simple Dijkstra algorithm and it works fine for little test cases. But it is not fast enough for random test.
Should I try Binary heap for Dijkstra or there is a better way?

Ok, i got it, thanks.

Python translation kumited.

approved.

i mean that i did not create any other accounts, but that I know 4 people who have solved this kata

That's fair enough, and that's why I tried clarifying what I meant in 2nd post.

Performance testing and because it is fun to have really big numbers.

...just because.

I meant: why do the inputs have to be this big?