Closed
Description
Bug report
Bug summary
When trying to plot a pandas Series with pyplot.bar(), it seems Series whose first index is not 0 cause a TypeError: only size-1 arrays can be converted to Python scalars exception when the bottom argument is given. When bottom is not given, bar width is not uniform as expected.
Code for reproduction
import pandas as pd
import matplotlib.pyplot as plt
df = pd.DataFrame({'i':[1,4,4,4,6,6], 'n':[3,5,1,2,3,4]})
ix = df.index%2 == 0
plt.figure()
# produces expected plot with bars offset from x axis
plt.bar(x=df.loc[ix,'i'], height=df.loc[ix, 'n'], bottom=df.loc[ix, 'i'])
plt.figure()
# produces expected plot
plt.bar(x=df.loc[ix,'i'], height=df.loc[ix, 'n'])
ix = df.index%2 == 1
plt.figure()
# the following fails with `TypeError`
# plt.bar(x=df.loc[ix,'i'], height=df.loc[ix, 'n'], bottom=df.loc[ix, 'i'])
# converting to a list fixes the issue and produces the expected outcome.
# note that casting both x and bottom to list is required to produce expected outcome.
plt.bar(x=list(df.loc[ix,'i']), height=df.loc[ix, 'n'], bottom=list(df.loc[ix, 'i']))
plt.figure()
# produces plot with unexpected bar width
plt.bar(x=df.loc[ix,'i'], height=df.loc[ix, 'n'])
# again, converting to list produces expected outcome
plt.figure()
plt.bar(x=list(df.loc[ix,'i']), height=df.loc[ix, 'n'])Matplotlib version
- Operating system: macOS 10.15
- Matplotlib version: 3.1.1
- Matplotlib backend (
print(matplotlib.get_backend())): MacOSX - Python version: 3.6.5
- Jupyter version (if applicable):
- Other libraries: pandas 0.25.2
Installed from conda, using default channel.