232 existing solutions make it unlikely there suddenly is a bug.

It is possible you're returning a wrong datatype, triggering weird messages. That would not be a bug in the tests; it could be considered a bug in your ( incomplete ) solution. Tests do not always handle that gracefully.

Hi! Try checking your solution againsta a dataset with multiple visits starting, ending or one starting and another ending at the exact same timestamp.

This KATA looks like cannot be solved by SQL at all because of this case "when no letter at all".

When I simply run the code "select * from letters", it gave me the below errors, which means it cannot be ran in SQL environment at all? Is it a bug?

undefined method `[]' for nil:NilClass
Simplified backtrace: block in run_test

Hi I have passed 56 test cases, but not sure there are 7 tests which always cannot pass.

It gave me below errors:

Test Failed
@@ ,when_happened ,visits_count
+++,2019-12-14 08:57:36 +0000,17
---,2019-12-11 08:57:36 +0000,33

or

Test Failed
@@,when_happened ,visits_count
→ ,2019-04-01 01:00:00 +0000,3→2

I am not able to see the full test cases so wondering what are the edge cases which blocked my passing?

I can run my code in local SQL client very well but in codewar, it always told me "undefined method `[]' for nil:NilClass Simplified backtrace: block (4 levels) in spec". Really don't understand what is the issue.

Me too. And I cannot use "With ... AS ()" subquery to escape the single quote, it will tell me the table not exists.