I like this, simpler than my solution!

it really should have a failing test on memory allocation with some IEnumberable the generates billions of values.

Reasons you may have been downvoted include:

1. This is a 7 kyu kata marked Fundamentals and Arrays; it is not marked for Optimization.

2. "Avoid premature optimization"; this code is very easy to understand and therefore should be very easy to maintain.

3. A good sorting algorithm will be O(n ln(n)) versus the O(n) of going through the list once. Yes, sorting is expected to take longer but is it significant?

I wrote a program that
(a) generates N random numbers,
(b) solves by a minimum selection method, and
(c) solves by sorting the array.

For N = 2,000,000 sorting the array was taking about 1/4 second (versus perhaps 30 ms for the other method). Unless I'm looking to run this a lot of times in an outer loop, that time isn't worth fussing over. AND, if I am running it many times I would first ask whether this is important for the overall program. If those 2,000,000 values are coming from a slow hard drive the program will appear slow with either algorithm.

If you're interested, at N = 20,000,000 the sorting took about 2 seconds (versus 80 ms). Here I would look at whether to optimize. Further, it took me a while to find my typo in the faster algorithm, which supports the maintainability aspect mentioned earlier.

aint no way💀

The second one was needed, as an error will be thrown without it

duplicate issue (see @donaldsebleung's issue below)

You get used to them when you start using them.

More code BUT definitely more optimal than 99% of the solutions here. Optimal is the way !!

Assuming that only one cell requires repositionion sort will take O(N)

agree. but better than n*n