You can see comments of a solution of a kata you haven't solved as long as you have the direct link to it, so they should be spoilered.

This IS the page for solutions, what are you talking about?

Closing as non-issue

if they hint/reveal the solution, yes!

The randomised test data can produce inputs with multiple periods with equal visit counts. This proposed solution should include v1.entry_time asc in the order by to pick the earlier event in that case.

Without the order by it is liable to pick the later group, and that's the wrong answer here (as when_happened should be the earliest timestamp).

Is there a need to mark comments to solutions as having spoilers?

could also use... the spoiler flag...

I maybe could, if you gave the full error message.

This is now failing some test cases?

more concise: sum((detail = 'good')::int) as good

hyper573's solution turbed out to be correct. I added a few test cases to weed out some incorrect solutions back in april.

hyper573's solution seemed like invalid on first look, but it turned out to be correct - i've updated my first comment back then.
His example does overlap, just not bidirectionally - thats how his solution works.
Kata's test cases were extended soon after that comment.

hyper573(4 kyu)2 months ago
I have the same issue as thurstonzd. I've found the answer wanted, but still don't get why it's accepted.

For example:

"2018-09-11 07:17:57" - "2018-09-14 07:17:57"

"2018-09-12 07:17:57" - "2018-09-16 07:17:57"

these visits do not overlap according to accepted conditions. Why?

@qgppl any update on when this will be fixed?