Fix: histogram sometimes does not obey selections

[JovanVeljanoski]

This issue could in partly due to count something being buggy when handling arrow data that has missing values. Not sure tho, but there are tests implemented that cover these things in both sides for safety (asserting the histogram output and the count output).

Workaround: If somebody is struggling with this issue, there is a simple workaround. Say column x is arrow column with missing values. All you need to do is:

df['x'] = df.x.as_numpy()

and everything should work as expected.

Checklist:

• [x] Make unit-tests exposing the issues
• [x] Make unit-tests pass
• [x] Party 🎉

[maartenbreddels]

Party 🎉

[JovanVeljanoski]

Party!! 🎉

Awesome, thank you!