Modify the legend of pandas bar plot

Question

I am always bothered when I make a bar plot with pandas and I want to change the names of the labels in the legend. Consider for instance the output of this code:

import pandas as pd
from matplotlib.pyplot import *

df = pd.DataFrame({'A':26, 'B':20}, index=['N'])
df.plot(kind='bar')

Now, if I want to change the name in the legend, I would usually try to do:

legend(['AAA', 'BBB'])

But I end up with this:

In fact, the first dashed line seems to correspond to an additional patch.

So I wonder if there is a simple trick here to change the labels, or do I need to plot each of the columns independently with matplotlib and set the labels myself. Thanks.

Answer 1

To change the labels for Pandas df.plot() use ax.legend([...]):

import pandas as pd
import matplotlib.pyplot as plt

fig, ax = plt.subplots()
df = pd.DataFrame({'A':26, 'B':20}, index=['N'])
df.plot(kind='bar', ax=ax)
#ax = df.plot(kind='bar') # "same" as above
ax.legend(["AAA", "BBB"]);

Another approach is to do the same by plt.legend([...]):

import matplotlib.pyplot as plt
df.plot(kind='bar')
plt.legend(["AAA", "BBB"]);