Oh, I got it now. You were right, I was couting rows and columns (row 0, row 1..., column 0, column 1). Thanks so much!

I explained bad myself, I think. Not talking of writting the grid but the position you should step. I mean, in the first example the first position (the starting "@") should be (1,2) (x,y), and the first step in order to go to the key is going up to (0,2), right? So when you say the array that should be returned with the pair x-y steps, you said "[(2, 1), (2, 0)...", there is where I am lost. Sorry for my clumsy english, I am explaining myself as hell...

Hello, sorry for the dumb-newbie question but I don't understand the first example. I mean, I suppose isn't wrong 'cause have been tested and achieved so some of them will pointed to it of ot were the case, but, "x" and "y" seems to be swapped to me. If "x" is for horizontal and "y" is for vertical: won't be "@" (1, 2) and first step to (0, 2), then (0, 1) and so on? Thanks.

Sure! Thanks a lot!

I know it has been more than a year, but just in case someone else has the same question. It depends on the kata creator, and most of the time, if validation is needed, it is specified in the description. However, in this specific case, it is not, and there are cases with negative numbers that need to be considered as 0, and values over 255 that need to be considered as 255. One thing that is specified is that you need 6 characters solution or won't work, this is because by default, zeroes on the left are deleted.
That's why it is useful to use control codes to identify when something is not working.

New thing learned! Achievement unlocked!

I was totally unaware of the existence of that method in Python.

I know, I know... I think I put some unnecessary parentheses extra...

That's the reason of my answer but I am doing my best. Anyway I asked this 4 months ago and now I understand it and noticed is pretty obvious. Thanks for the answer.

I see, thanks for the clarification, and sorry tagging it as 'spoiler', didn't know about notifications, I'll remember in future cases.

As I had no answer in almost a month I will suppose issue can be marked as solved.

There seems to be a problem on every last test of each block. Tried with many different codes and none of them work with mentioned tests but works with others. Here is an example of failed test:
Log [5, 3, 1, 8] 1 [5, 1] 6 should equal 13
If you pay attention there isn't a combination that sums 13 and I am deleting correct positions before sum the others, which result is 6. I may be wrong or missing something but anyway it is weird the fail of last tests per block overall if the other ones are working. Can you check it? Thanks!

And it didn't. It seems something implemented in 3.10 according to the other commentary. I used this in C# and Java, but didn't tried in Python just thinking it hadn't it.