Repository navigation
[Bug]: plt.bar legend displays incorrect style for empty data #21506
Description
Bug summary
Legend items created from empty plt/ax.bar calls are displayed with the default style and the first color of the theme's cycle. This behavior contradicts the behavior of plt/ax.plot and plt/ax.scatter.
Code for reproduction
import matplotlib.pyplot as plt
# The bug happens with or without this import.
import numpy as np# Actual outcome code
plt.bar([], [], alpha=0.3, label="[], [], alpha=0.3")
plt.bar([], [], alpha=0.3, label="[], [], alpha=0.3")
plt.bar([np.nan], [np.nan], alpha=0.3, label="[np.nan], [np.nan], alpha=0.3")
plt.bar([], [], alpha=0.3, label="[], [], alpha=0.3")
plt.bar([1], [1], alpha=0.3, label="[1], [1], alpha=0.3")
plt.legend()# Expected outcome code
plt.plot([], [], alpha=0.5, label="[], [], alpha=0.5")
plt.plot([], [], alpha=0.5, label="[], [], alpha=0.5")
plt.plot([np.nan], [np.nan], alpha=0.5, label="[np.nan], [np.nan], alpha=0.5")
plt.plot([], [], alpha=0.5, label="[], [], alpha=0.5")
plt.plot([1], [1], alpha=0.5, label="[1], [1], alpha=0.5")
plt.legend()Actual outcome
Color cycling and alpha are not respected for the plt.bar([], [], ...) calls.
Expected outcome
Color cycling and alpha are respected for the plt.plot([], [], ...) calls.
Operating system
Linux
Matplotlib Version
3.2.2
Matplotlib Backend
module://ipykernel.pylab.backend_inline
Python version
3.7.12 (default, Sep 10 2021, 00:21:48) [GCC 7.5.0]
Jupyter version
Colab
Other libraries
No response
Installation
No response
Conda channel
No response
Reacted by Kevin Max
Activity
Fixed bug with legend color and empty bars. Closes matplotlib#21506.
Fixed bug with legend color and empty bars. Closes matplotlib#21506.
Fixed bug with legend color and empty bars. Closes matplotlib#21506.
Fixed bug with legend color and empty bars. Closes matplotlib#21506.
kma-code commentedon Apr 3, 2025
Bug persists in Matplotlib 3.3.4.
If you are trying to generate a legend using an empty plt.bar, a workaround is to plot a bar with some data and set width = 0:
Operating system
Mac OS
Matplotlib Version
3.2.4
Matplotlib Backend
module://ipykernel.pylab.backend_inline
Python version
3.11
Jupyter version
IPython : 8.26.0
ipykernel : 6.29.5
ipywidgets : not installed
jupyter_client : 8.6.2
jupyter_core : 5.7.2
jupyter_server : 2.15.0