Use subqueries in Flop.count for queries with group_by

[michallepicki]

See also related discussion

[woylie]

Nice, thanks, we just need some tests 👍

[codecov]

Codecov Report

Patch coverage: 100.00% and project coverage change: +0.09 🎉

Comparison is base (f587432) 88.29% compared to head (5d62388) 88.38%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #336      +/-   ##
==========================================
+ Coverage   88.29%   88.38%   +0.09%     
==========================================
  Files          13       13              
  Lines         880      887       +7     
==========================================
+ Hits          777      784       +7     
  Misses        103      103              

Impacted Files	Coverage Δ
lib/flop.ex	94.70% <100.00%> (+0.12%)	⬆️

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[michallepicki]

Looks like we don't need to wrap queries with distinct, from Ecto.Repo.aggregate docs:

If the query has a limit, offset, distinct or combination set, it will be automatically wrapped in a subquery in order to return the proper result.

Maybe this was different in the past and that's why Scrivener had a special case for this. I'll remove this part.

[woylie]

Thank you!