python - How to return a matplotlib.figure.Figure object from Pandas plot function

Question

I have a dt:

>>> dt
            sales                mg         
ID         600519   600809   600519   600809
RPT_Date                                    
20060331  13.5301   5.8951   9.4971   3.0408
20060630   6.6048   2.4081   4.3088   1.4040
20060930  12.3889   3.6053   9.1455   2.0754
20061231  16.5100   3.3659  12.4682   1.8810
20070331  15.8754   5.9129  11.5833   3.6736
20070630  10.4155   3.5759   7.8966   2.2812
20070930  18.2929   3.5280  14.3552   2.1584
20071231  27.7905   5.4510  23.7820   3.2568

And use pandas function plot to create a barplot object

>>> fig = dt.plot(kind='bar', use_index=True)
>>> fig
<matplotlib.axes.AxesSubplot object at 0x0B387150>

But I want a <matplotlib.figure.Figure object> instead to pass to other function, just like the object type below:

>>> plt.figure()
<matplotlib.figure.Figure object at 0x123A6910>

So how can I transform <matplotlib.axes.AxesSubplot object> to <matplotlib.figure.Figure object> or directly return "Figure object" from Pandas plot ?

Answer 1

You can get the figure from the axes object:

ax.get_figure()

Full example:

df = pd.DataFrame({'a': range(10)})
ax = df.plot.barh()
ax.get_figure()