axvspan() sometimes breaks when using matplotlib wrappers

[darothen]

Hi,

I was trying to highlight some data in a DataFrame using the axvspan method from matplotlib and I ran into some ill behavior. Here's a minimal example:

import numpy as np
import pandas as pd
import matplotlib.dates as mdates

from pylab import *
ion()

dates = pd.date_range(start="10/01/2013", end="10/04/2013", freq='H')
data = np.random.random_sample(len(dates))

my_df = pd.DataFrame({'r': data}, index=dates)

figure()
my_df.plot()
axvspan(*mdates.datestr2num(["10/02/2013", "10/03/2013"]),
             facecolor='red', alpha=0.5)

This produces the following figure:

However, when I instead invoke the plotter via my_df.plot(x_compat=True), things are kosher:

save for the fact that the advanced x-axis formatting has been lost (easy enough to replace). I'm not sure if this is an unintended feature or a bug, but it's not documented anywhere.

[selasley]

I had a similar problem recently and found that passing datetimes into axvspan works. So

axvspan(*pd.to_datetime(["10/02/2013", "10/03/2013"]),
facecolor='red', alpha=0.5)

works with my_df.plot() and my_df.plot(x_compat=True)

[jreback]

closing as not a bug

[michaelaye]

This should be advertised better, extremely important feature!