I didn't know that cause that's not what you said in your comment. That's still slower for large number sizes, e.g 32 bit integers (the most common size for an integer in cpp), although if we're being specific since we're dealing with signed integers here, so the max would be like 2^31 - 1. I agree that for small numbers, the iterative method is faster, but not by much as both methods are practically instant when dealing with small numbers. It's better to make the function scalable to larger numbers by using this method.

no, this solution about a lot of copiyng

too much copiyng

WTF

It's funny because this was actually pretty much my original answer, except that I didn't account for numbers less than zero.
The error message wasn't very helpful, so I found myself jumping through hoops trying to figure out how to handle floating point inaccuracies.

I feel smartn't seeing this

Seemed like a simple task, but I broke my head solving it, 2 out of 3 conditions in the task were met and the last one was not accepted by me. Of course, I had to resort to help....... It turned out to be the same thing I wrote, just with some minor changes, which also turned out to be tricky, but now I see that it's even easier. I'm going to be a wreck today, work in 4 hours and I haven't slept.....

Looks very simple, clear, and easy. Thank you! As for me: as always I created more complicated solution :)

std::map is actually O(logn) to find. A hashmap (std::unordered_map) would be what to use for efficiency

Fantastic. I was trying to create this simple O(n) O(n) solution. Only when you see it do you realise how simple it is. Kicking myself, as I was almost there twice but deleted and went down other avenues.
Great work.