BUG: hidden ticklabels with sharex and secondary

[TomAugspurger]

Closes #9158

import pandas as pd
d = {'A' : [1., 2., 3., 4.], 'B' : [4., 3., 2., 1.], 'C': [5, 1, 3, 4]}
df = pd.DataFrame(d, index=pd.date_range('2014 10 11', '2014 10 14'))

axes = df[['A','B']].plot(subplots=True)
df['C'].plot(ax=axes[0], secondary_y=True)   # x tick labels dissappear when this line is executed

Introduced in the subplot refactoring in #7457

cc @sinhrks

The fix is pretty simple, but I'm worried I may have broken other things. No tests failed though, so hopefully we're good.

[TomAugspurger]

My reformatting hides the change.

The only difference is in the call to _handle_shared_axes, naxes is now nrows * ncols, instead of len(all_axes). The problem was that all_axes is 3 now, since there's the original two subplots and the additional 3rd plot, which is getting plotted on the first subplot. This confused _handle_shared_axes.

[jreback]

no spaces needed on multiple lines (it will auto-wrap)

[jreback]

looks good. minor doc issue.

[jreback]

@TomAugspurger looks ok, pls rebase.

[jreback]

needs a rebase

[TomAugspurger]

Rebased... will merged later today.

[TomAugspurger]

The failure here was a build failure on a job that doesn't run the plotting code. Going to merge

[TomAugspurger]

Well, I messed that up. Pushed before --continuing my rebase, and it looks like github has it's done with this PR

[shadowleaves]

this problem seems to be still existing in pandas 0.17.0, on subplots with at least two rows:

import numpy as np
import pandas as pd
from matplotlib import pyplot as plt

mat = np.random.rand(10,2)
df = pd.DataFrame(mat, columns=['A','B'])

fig, axes = plt.subplots(nrows=2, ncols=2)                                 
df.plot(secondary_y='B', ax=axes[0, 1], legend=False)