No chromosome is written like this, X comes first then Y.

Alternatively you can just include an index in the table and limit the recursion with "where index <= 90"

if the condition to check will be "YX" , i think this solution wouldnt work ..
Traceback (most recent call last):
File "/workspace/default/.venv/lib/python3.10/site-packages/codewars_test/test_framework.py", line 112, in wrapper
func()
File "/workspace/default/tests.py", line 8, in basic_tests
test.assert_equals(chromosome_check('YX'), 'Congratulations! You're going to have a son.')
File "/workspace/default/solution.py", line 4, in chromosome_check
return "Congratulations! You're going to have a {}.".format(gender[sperm])
KeyError: 'YX'

It's a thing that only works for simple queries.

The docs says:

This works because PostgreSQL's implementation evaluates only as many rows of a WITH query as are actually fetched by the parent query.
Using this trick in production is not recommended, because other systems might work differently.
Also, it usually won't work if you make the outer query sort the recursive query's results or join them to some other table,
because in such cases the outer query will usually try to fetch all of the WITH query's output anyway.

https://www.postgresql.org/docs/13/queries-with.html

Yes, this is wrong according to the program specification—that's not being slightly picky at all! It's a common error class that you need to watch out for because it can have real consequences, for example if you set about building a fence and find yourself one post short:

https://en.wikipedia.org/wiki/Off-by-one_error

Could someone explain to me how this works in steps??? thank you so much!

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).

This is now failing some test cases?

int / int = int
5 / 4 = 1, not 1.250000

Im felling very dumby

R maintains the letters vector, so no need to write out entire alphabet.

I like how concise this is, but I'm not sure an order by col index is a best practice. Not that I don't also use it, but that's mostly when I'm in an RDBMS that won't let me reference a derived field by its alias. PostGres does allow it (ie: order by batting_average desc).

I need to get more accustomed to the ::casting operators in postgres. Oracle and SQLServer don't have them.