Fix MPL dt conversion (epoch2num deprecation) #34850

[misantroop]

Fix datetime conversion after matplotlib (3.3.0 update) deprecating epoch2num.

• closes BUG: epoch2num matplotlib deprecation warning #34850
• tests added / passed
• passes black pandas
• passes git diff upstream/master -u -- "*.py" | flake8 --diff
• whatsnew entry

[TomAugspurger]

Thanks @misantroop. Seems to be some test failures. Can you check?

____________________ TestDateTimeConverter.test_conversion _____________________
[gw2] darwin -- Python 3.6.10 /Users/runner/miniconda3/envs/pandas-dev/bin/python

self = <pandas.tests.plotting.test_converter.TestDateTimeConverter object at 0x161aaac18>

    def test_conversion(self):
        rs = self.dtc.convert(["2012-1-1"], None, None)[0]
        xp = datetime(2012, 1, 1).toordinal()
>       assert rs == xp
E       assert 15340.0 == 734503

pandas/tests/plotting/test_converter.py:150: AssertionError

Can you post images of the outputs with / without your changes for matplotlib 3.3 and an earlier version?

[igavronski]

Please see an alternative circumvent that takes into account the same problem here:
#34850 (comment)

[misantroop]

Issue fixed in MPL: matplotlib/matplotlib#17983