fix: Fix boolean distinct

[ritchie46]

No description provided.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 81.32%. Comparing base (2398b47) to head (dc99ab9).
Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main   #16765   +/-   ##
=======================================
  Coverage   81.32%   81.32%           
=======================================
  Files        1424     1424           
  Lines      187215   187223    +8     
  Branches     2726     2726           
=======================================
+ Hits       152250   152267   +17     
+ Misses      34468    34459    -9     
  Partials      497      497           

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

[codspeed-hq]

CodSpeed Performance Report

Merging #16765 will improve performances by 20.14%

_{Comparing null_count (dc99ab9) with main (2398b47)}

Summary

⚡ 1 improvements
✅ 36 untouched benchmarks

Benchmarks breakdown

	Benchmark	main	null_count	Change
⚡	test_groupby_h2oai_q1	2.7 ms	2.3 ms	+20.14%

[henrikig]

@ritchie46 Something seems off in this else branch, and I'm not getting the numbers I'd expect when trying to use this as you advised in #16588.

Take for instance the case when you have an array with null+true or null+false, you will never get a number less than 2 with this logic, while it should actually be 1 if I understand the function signature correctly.

Also, we're using the unset_bits() from the values directly when we know that the array has nulls in it, which means the bits could be anything if I understand it correctly.

Some examples of what I get, and what I would expect:

[True, None, True, None, None] # got 3, expected 1
[False, None, False, None, None] # got 2, expected 1
[True, False, None, False, None, None] # got 3, expected 2

Added a debug_assert_eq in my code as well as below, which fails:

    let num_unique = usize::from(n_true > 0) + usize::from(n_false > 0);
    let distinct_count = arr.distinct_count();
    debug_assert_eq!(num_unique, distinct_count);

[henrikig]

Made a new PR, see #16780