docs: Fix docstrings for ATAN2 and ATAN2D SQL functions

[janscholten]

The docs for ATAN2 and ATAN2D (SQL functions) were incorrect because they claimed to be dividing arg2/arg1 where it is actually the other way around.

Alternatively, one could swap the arguments in the implementation, but I like the implementation and it follows numpy.

Also, I changed the example in the docstring so that it uses lengths of 0, 1, or 2, and not the ambiguous "lengths" of 90, 360 etc.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 80.21%. Comparing base (8a76dad) to head (50ff1fe).
Report is 47 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main   #19351      +/-   ##
==========================================
- Coverage   80.24%   80.21%   -0.03%     
==========================================
  Files        1523     1523              
  Lines      209545   209795     +250     
  Branches     2434     2434              
==========================================
+ Hits       168148   168294     +146     
- Misses      40842    40946     +104     
  Partials      555      555              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[orlp]

We should be consistent with the other SQL implementations. I believe most follow the convention atan2(y, x) ~ atan(y / x), so we should do that as well.

[janscholten]

We should be consistent with the other SQL implementations. I believe most follow the convention atan2(y, x) ~ atan(y / x), so we should do that as well.

Agreed.