ENH: Allow subplots axes to be passed into DataFrame.plot()

[moorepants]

This works nicely:

Python 2.7.5+ (default, Sep 19 2013, 13:48:49) 
Type "copyright", "credits" or "license" for more information.

IPython 1.1.0 -- An enhanced Interactive Python.
?         -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help      -> Python's own help system.
object?   -> Details about 'object', use 'object??' for extra details.

In [1]: import pandas, numpy

In [2]: import matplotlib.pyplot as plt

In [3]: df1 = pandas.DataFrame(numpy.random.random((5, 4)))

In [4]: df2 = pandas.DataFrame(numpy.random.random((5, 4)))

In [5]: ax = df1.plot()

In [6]: df2.plot(ax=ax)
Out[6]: <matplotlib.axes.AxesSubplot at 0x3f4f910>

In [7]: plt.show()

But this doesn't:

In [8]: ax = df1.plot(subplots=True)

In [9]: df2.plot(subplots=True, ax=ax)
---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-9-70822353a5d1> in <module>()
----> 1 df2.plot(subplots=True, ax=ax)

/usr/local/lib/python2.7/dist-packages/pandas/tools/plotting.pyc in plot_frame(frame, x, y, subplots, sharex, sharey, use_index, figsize, grid, legend, rot, ax, style, title, xlim, ylim, logx, logy, xticks, yticks, kind, sort_columns, fontsize, secondary_y, **kwds)
   1634                      logy=logy, sort_columns=sort_columns,
   1635                      secondary_y=secondary_y, **kwds)
-> 1636     plot_obj.generate()
   1637     plot_obj.draw()
   1638     if subplots:

/usr/local/lib/python2.7/dist-packages/pandas/tools/plotting.pyc in generate(self)
    853         self._args_adjust()
    854         self._compute_plot_data()
--> 855         self._setup_subplots()
    856         self._make_plot()
    857         self._post_plot_logic()

/usr/local/lib/python2.7/dist-packages/pandas/tools/plotting.pyc in _setup_subplots(self)
    909                 ax = self._maybe_right_yaxis(ax)
    910             else:
--> 911                 fig = self.ax.get_figure()
    912                 if self.figsize is not None:
    913                     fig.set_size_inches(self.figsize)

AttributeError: 'numpy.ndarray' object has no attribute 'get_figure'

It would be nice if you could use the subplots flag, because it makes it easier to compare two data frames visually.

[TomAugspurger]

This seems a bit tricky API wise. We'd need to figure out what to do if df2 and df1 are not indexed the same.

What would you think about having more control over the layout if you had a MultiIndex for the columns?

df3 = pd.concat([df1, df2], axis=1, keys=['a', 'b'])
In [1]: df3
Out[68]: 
          a                                       b                      \
          0         1         2         3         0         1         2   
0  0.403222  0.179409  0.819205  0.839402  0.421360  0.334930  0.251081   
1  0.689900  0.969358  0.247526  0.670381  0.509042  0.237918  0.794933   
2  0.998004  0.965907  0.354623  0.109929  0.160774  0.130975  0.591787   
3  0.761342  0.420306  0.571551  0.109936  0.581581  0.925967  0.586253   
4  0.750932  0.624856  0.975460  0.860216  0.414069  0.883192  0.668196   


          3  
0  0.952108  
1  0.420736  
2  0.378974  
3  0.037831  
4  0.706987  

[5 rows x 8 columns]

And then you could to subplots by a level of the MultiIndex? That removes any ambiguity of differently index df1 and df2.

---

as a workaround for now:

axes = df1.plot(subplots=True)

for i, ax in enumerate(axes):
    df2[i].plot(ax=ax)

plt.draw()

[moorepants]

This works:

axes = df1.plot(subplots=True)

for i, ax in enumerate(axes):
    df2.iloc[:, i].plot(ax=ax)

plt.draw()

The multiindex idea sounds good too.

[jreback]

#6667 will make this possible I think

[jreback]

dupe of #6667