Bug summary

ax.bar_label appears not to be robust to bars with missing (nan) values when also including error values. This issue is similar to #20058, but occurs in each of three cases:

Case 1. When a dependent value is missing.
Case 2. When an error value is missing.
Case 3. When both a dependent value and an error value are missing.

The error seems to happen here, but I don't know the code well enough to pinpoint what should change to fix this:

Code for reproduction

#%% Case 1: Missing dependent value
import matplotlib.pyplot as plt
import numpy as np
ax = plt.gca()
bars = ax.bar([0, 1, 2], [np.nan, 0.3, 0.4], yerr=[1, 0.1, 0.1])
ax.bar_label(bars)

#%% Case 2: Missing error value
import matplotlib.pyplot as plt
import numpy as np
ax = plt.gca()
bars = ax.bar([0, 1, 2], [0, 0.3, 0.4], yerr=[np.nan, 0.1, 0.1])
ax.bar_label(bars)

#%% Case 3: Missing dependent and error values
import matplotlib.pyplot as plt
import numpy as np
ax = plt.gca()
bars = ax.bar([0, 1, 2], [np.nan, 0.3, 0.4], yerr=[np.nan, 0.1, 0.1])
ax.bar_label(bars)

Actual outcome

runcell('Case 3: Missing dependent and error values', 'C:/Users/jam/Documents/GitHub/ci-greedy-agents-base/untitled2.py')
Traceback (most recent call last):

File "C:\ProgramData\Miniconda3\lib\site-packages\spyder_kernels\py3compat.py", line 356, in compat_exec
exec(code, globals, locals)

File "c:\users\jam\documents\github\ci-greedy-agents-base\untitled2.py", line 27, in
ax.bar_label(bars)

File "C:\ProgramData\Miniconda3\lib\site-packages\matplotlib\axes_axes.py", line 2641, in bar_label
endpt = err[:, 1].max() if dat >= 0 else err[:, 1].min()

IndexError: too many indices for array: array is 1-dimensional, but 2 were indexed

Expected outcome

Maybe either raise an error telling me what I should do instead, or have the code resolve whatever the source is on the backend? Ideally, I think the following should happen:

Case 1. Raise an error that there is no value to apply the errorbar value to.
Cases 2 & 3. Ignore the missing value and move on to the next.

Additional information

No response

Operating system

Windows 10.1

Matplotlib Version

3.5.1

Matplotlib Backend

module://matplotlib_inline.backend_inline

Python version

3.9.5

Jupyter version

Spyder 5.3.0

Installation

conda