Indeed, but this is SQL only kata, so it shows that you can done some think on the database level.

Agreed, to be honest, I've checked the test and don't know why they are not valid, they seem to match the rules. And I'm saying from logical point of view, not regexs

Fixed

I've done that before, but then when there is for exempla 3.40 in the result table it shows 3.4 . Maybe it is the problem of the seqel gem? But it is not preserving exact 2 digits

I'm not sure if that was it, but there was some example tests in the kata, because i did't provide any. And there was not connected with the kata - Generic tests for items.
Can you confirm that? I've added example tests for this kata. Should be fine now

I've fixed it. And change to order by price_per_kg then by name, makes more sense according to title

Sorry for that, when I was fixing test, there was some old cache in the Kata editor, and I didn't notice that. I've already fixed your remarks.

I've added additional description, and I've formatted text to be more clear

I've added additional description and additional sorting. Indeed I've missed that out:)

Ok, fixed, now it is sorting results in tests, so user does not need to care about ordering

Yes it is but postgres does not guarantee ordering from what I have read, but I'm not and expert in that matter, I will add sort in the test to preserve solutions. Next time I will add ordering in the problem description

Ok, I've check the docs, and the order is indeed not guaranteed - I just assumed that without ORDER BY is for lower kyu,i think that in most 8kyu and 7kyu there is no ORDER BY.
If there are already some people who solve the proble, can I change the solution by adding SORT BY ?