Plotting ExtensionArrays

[andrewgsavage]

• closes Plotting ExtensionArrays #26173
• tests added / passed
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

How do I go about adding tests for that rely on an external module?
Would be something like

df = pd.DataFrame({
        "length" : pd.Series([1,2],dtype = pd.Int16Dtype),
        "width" : PA_([2,3], dtype="pint[m]"),
        "distance" : PA_([2,3], dtype="in"),
        "height" : PA_([2,3], dtype=ureg.m),
        "depth" : PA_.from_1darray_quantity(Q_([2,3],ureg.mm)),
    })
df.plot()

[pep8speaks]

Hello @andrewgsavage! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2019-06-30 07:12:34 UTC

[jreback]

rather than this i would fix select_dtypes to understand the is_numeric flag of the EA dtype; it’s possible that actually need something more because for example datetimes are not numeric but are allowed for plotting

[TomAugspurger]

Agreed that this should be addressed in select_dtypes rather than here.

[andrewgsavage]

I'm a bit confused here, df.select_dtypes('number') will not return columns with bool dtype. However is_numeric_dtype(dtype) will return True for bool. Should is_numeric_dtype be used in select_dtypes?

[TomAugspurger]

I don't quite understand this. Are you unconditionally adding adding an EA if it's considered numeric here?

[jreback]

can you merge master

[WillAyd]

Nice idea but looks like this has gone stale and I think has a few dependencies that need to be cleaned up. In any case @andrewgsavage if interested in continuing please ping and can reopen