BUG: Handle line and bar on the same plot

[MartinBraquet]

• closes BUG: fails to plot 2 plots with secondary y axis when index type is PeriodIndex and plot kinds are different #61161
• Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• Added type annotations to new arguments/methods/functions.
• Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

This code

import pandas as pd
import matplotlib.pyplot as plt

index = pd.period_range('2023', periods=3, freq='Y')
df = pd.DataFrame({
    'col1': [10, 20, 30],
    'col2': [40, 25, 10]
}, index=index)

fig, ax = plt.subplots()
df['col1'].plot(kind='bar', ax=ax)
ax2 = ax.twinx()
df['col2'].plot(kind='line', ax=ax2, color = 'r')
plt.show()

from the linked issue gives this plot:

And swapping the lines (line before bar)

df['col2'].plot(kind='line', ax=ax2, color = 'r')
df['col1'].plot(kind='bar', ax=ax)

gives

I believe the difference of results stems from the fact that the figure is always cropped to match the x-range of the last plot. We can certainly adjust the PR to make the first figure behave like the second one if desired.

Please provide feedback on the overall approach. Once confirmed, I'll expand and cover with tests.

[rhshadrach]

Thanks for the PR! Can you add a test in pandas/tests/plotting/test_series.py. Make the plots and check that the x-axis is as expected.

[rhshadrach]

Why is the modification of freq necessary? I worry that we don't own ax, and that we shouldn't be mutating it. But I'm not too familiar with the plotting code; do we do this anywhere else? I don't see any instances at a glance.

[MartinBraquet]

The only reason is to pass the freq to PeriodConverter, which currently only accepts ax as input arg, not freq. It can obviously be made cleaner. I meant to modify PeriodConverter or find a cleaner way than hacking around ax.

I'd just like to get a thorough review of the overall approach, i.e. shifting the tick positions, as it would apply to all box plots with PeriodIndex, and hence is a potential major change. Not sure how well it is covered by tests currently; I don't want to insert regressions or breaking changes. It would be good to have it first checked by a maintainer well-versed with matplotlib, as I'm not so familiar with all the idiosyncrasies of the ax object.