BUG: Blank axes when mixing pandas- and matplotlib-style plots in matplotlib>=3.4.0

[kandersolar]

Pandas version checks

• I have checked that this issue has not already been reported.

• I have confirmed this bug exists on the latest version of pandas.

• I have confirmed this bug exists on the master branch of pandas.

Reproducible Example

import matplotlib
import matplotlib.pyplot as plt
import pandas as pd

tz = 'Etc/GMT+5'  # change to None
freq = 'h'  # change to 'd'
times = pd.date_range('2019-01-01', freq=freq, periods=3, tz=tz)
series = pd.Series(range(len(times)), index=times)

fig, ax = plt.subplots()
series.plot(ax=ax)
ax.errorbar([times[1]], [1], yerr=[[1], [2]], c='k')
plt.title(f"freq='{freq}' tz={tz}\n"
          f'matplotlib=={matplotlib.__version__} pandas=={pd.__version__}')

Issue Description

For timeseries data, mixed pandas-style (Series.plot(ax=ax)) and pyplot-style (ax.plot()) plotting has different behavior starting in matplotlib==3.4.0. I can't make sense of what's going on, but the difference depends on whether the Series is tz-localized and the frequency of the data. Here are some comparisons using the MRE:

Description	matplotlib==3.3.4	matplotlib==3.4.0
Axes show no lines at all
Axes show line from Series.plot(), but not the ax.errorbar
Both lines display, though the error bar is timeshifted
No apparent difference

Expected Behavior

I suppose I expect to be able to mix and match pandas- and matplotlib-style plotting calls like I can with matplotlib 3.3.4.

Installed Versions

INSTALLED VERSIONS

commit : 66e3805
python : 3.7.7.final.0
python-bits : 64
OS : Windows
OS-release : 10
Version : 10.0.19041
machine : AMD64
processor : Intel64 Family 6 Model 158 Stepping 10, GenuineIntel
byteorder : little
LC_ALL : None
LANG : None
LOCALE : None.None

pandas : 1.3.5
numpy : 1.19.5
pytz : 2019.3
dateutil : 2.8.1
pip : 21.3.1
setuptools : 50.3.0
Cython : 0.29.26
pytest : 6.2.3
hypothesis : None
sphinx : 3.2.0
blosc : None
feather : None
xlsxwriter : None
lxml.etree : None
html5lib : 1.0.1
pymysql : None
psycopg2 : None
jinja2 : 2.11.3
IPython : 7.31.0
pandas_datareader: None
bs4 : None
bottleneck : None
fsspec : None
fastparquet : None
gcsfs : None
matplotlib : 3.4.0
numexpr : 2.7.3
odfpy : None
openpyxl : None
pandas_gbq : None
pyarrow : None
pyxlsb : None
s3fs : None
scipy : 1.5.4
sqlalchemy : None
tables : 3.6.1
tabulate : None
xarray : None
xlrd : None
xlwt : None
numba : 0.53.1

[jorisvandenbossche]

@kanderso-nrel thanks for the report! I can confirm this still happens on the main branch of pandas as well, and also with the latest matplotlib 3.5.1

cc @QuLogic in case you are aware of similar issues / reports at the matplotlib side, or know what might have caused this (didn't yet further investigate it myself)

[QuLogic]

Possibly the Pandas date converters auto-registration? 3.3 changed the epoch of dates, I don't know if anything was in 3.4.