BUG: axvspan breaks when using pandas .plot() wrapper

[DanielHabenicht]

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 main branch of pandas.

Reproducible Example

df = pd.read_csv("export.csv", index_col="Timestamp", parse_dates=True)

start = pd.Timestamp("2019-01-15 23:00:00+0000", tz="UTC")
end = pd.Timestamp("2019-02-01 09:50:00+0000", tz="UTC")
print(start)
print(df.index[0])
print(end)

fig, ax = plt.subplots()
ax.axvspan(
    start, # or df.index[0]
    end, # or df.index[-1]
    color="red",
    alpha=0.1,
    lw=0,
    zorder=1,
)

# Using the `ax.plot` method works
# ax.plot(df2.index, df2["n1"], color="black", alpha=0.2, linestyle="solid", zorder=3)

df.plot(ax=ax, alpha=0.2, linestyle="solid", zorder=3)

# Using `axvspan` before and after the plot actually works too
# Only using this axvspan method results in a blank graph
# ax.axvspan(
#     start,
#     end,
#     color="red",
#     alpha=0.1,
#     lw=0,
#     zorder=1,
# )

plt.show()

Issue Description

Using the wrapper functions somehow generates weird behaviour for plotting.
Interestingly only with the data I used here: export.csv

Using the axvspan before df.plot():

Using the axvspan after df.plot():

Expected Behavior

Using axvspan before and after df.plot or using the standard ax.plot(x,y) function:

Related

There was some weirdness reported before; #5092

Installed Versions

INSTALLED VERSIONS

commit : 2e218d1
python : 3.10.10.final.0
python-bits : 64
OS : Linux
OS-release : 5.15.90.1-microsoft-standard-WSL2
Version : #1 SMP Fri Jan 27 02:56:13 UTC 2023
machine : x86_64
processor :
byteorder : little
LC_ALL : None
LANG : C.UTF-8
LOCALE : en_US.UTF-8

pandas : 1.5.3
numpy : 1.24.2
pytz : 2022.7.1
dateutil : 2.8.2
setuptools : 67.4.0
pip : 23.0
Cython : None
pytest : 7.2.1
hypothesis : None
...
xlrd : None
xlwt : None
zstandard : None
tzdata : None

[ritikBhandari]

Hi @DanielHabenicht ,

When I run the code, I got the same result as it was expected to show for using axvspan() after df.plot().

Regarding why the stuff works different when applied to different position is when you call df.plot(), it creates a new plot on the current figure with a new set of axes(). When we axvspan() after that, it is applied to the current set of axes which by default is the last set of axes created by the most recent plot command. So when you call axvspan() after df.plot(), it applies the shaded region to the plot that was just created by df.plot().

On the other hand, when you call axvspan() before it, the shaded region is created on the plot before df.plot() is called. So it is drawn behind the plot.

You can play around with zorder parameter to make sure it is shown in front of the plot.

[DanielHabenicht]

Hi @ritikBhandari, thanks for taking the time to reproduce the issue.

I did try it with zorder, as you can see in my example. It did not change the result:

• If I apply axvspan only before df.plot only the graph shows up

• If I apply axvspan only after df.plot nothing shows up.

• If I apply it at both positions or use ax.plot it works as expected.

So in short, you can't reproduce it? Then I will try to reproduce it on another machine.

One more question:
For df.plot(ax=ax) I was under the impression that it would reuse my supplied axis, because I don't see why it is supplied otherwise?

[ritikBhandari]

So I tried to reproduce the code and it is showing me the shades when I apply axvspan after df.plot. This issue is mostly because of the machine you are using I guess.

And regarding why it isn't applying shades when used before df.plot, it is because matplotlib by-default covers the first layer (which in our case is the shade) when it is covered by another layer(our actual plot). So that is the reason it doesn't work when applied before df.plot.

[DanielHabenicht]

Alright, seems like I also can't reproduce the df.plot() => axvspan case on a different machine.

But I would still expect that:

ax.axvspan(start, end)
ax.plot(df.index, df["n1"], linestyle="solid", zorder=3)

behaves the same as:

ax.axvspan(start, end)
df.plot(ax=ax, linestyle="solid", zorder=3)

[azjps]

I didn't reproduce the exact same issue but I'm fairly certain this is the same underlying issue as #52895; df.plot() & axvspan() look as expected after ignoring inferred_freq

[mroeschke]

Closing as a similar issue to #52895