NaT in date columns make histograme go crazy

Hello,

I am referencing issue 111, which happens again in version 2.8.0

import pandas as pd
import pandas_profiling
import numpy as np


def test_issue_111():
    years = pd.Series(["{:02d}/01/{}".format(x, y) for x in range(1989, 2020) for y in range(1, 30)])

    df = pd.DataFrame(np.random.randint(0, len(years), size=(len(years), 4)), columns=list('ABCD'))
    df['E'] = pd.to_datetime(years)
    df['E'].iloc[10:100] = pd.NaT

    report = df.profile_report()
    report.to_file('test.html')

_From the post by @sbrugman in https://github.com/pandas-profiling/pandas-profiling/issues/111#issuecomment-497034063_

[mbh86]

@sbrugman, I tested it and it works, thank you.

[sbrugman]

The v2.9.0rc1 release is out, and should resolve this issue. Until this version is fully released, you can install it via pip in the following way:

pip install --pre -U pandas-profiling

It would be very helpful to know if the release candidate adequately solves the issue.