bar_label fails with nan data values #20058
Description
Bug report
Bug summary
ax.bar_label is not robust to bars with missing (nan) data values.
The problem is in here:
matplotlib/lib/matplotlib/axes/_axes.py
Lines 2625 to 2635 in 7a12f32
Because nan is neither greater than nor less than 0, ha and va never get defined, and don't have default values.
Code for reproduction
ax = plt.gca()
bars = ax.bar([0, 1], [np.nan, 1])
ax.bar_label(bars)Actual outcome
---------------------------------------------------------------------------
UnboundLocalError Traceback (most recent call last)
<ipython-input-65-f996ada4311b> in <module>
1 ax = plt.gca()
2 bars = ax.bar([0, 1], [np.nan, 1])
----> 3 ax.bar_label(bars)
~/miniconda3/envs/py39/lib/python3.9/site-packages/matplotlib/axes/_axes.py in bar_label(self, container, labels, fmt, label_type, padding, **kwargs)
2636 annotation = self.annotate(fmt % value if lbl is None else lbl,
2637 xy, xytext, textcoords="offset points",
-> 2638 ha=ha, va=va, **kwargs)
2639 annotations.append(annotation)
2640
UnboundLocalError: local variable 'ha' referenced before assignmentExpected outcome
No error; no annotation generated. (Or maybe empty string?)
Matplotlib version
- Operating system: macos
- Matplotlib version (
import matplotlib; print(matplotlib.__version__)): 3.4.1 - Matplotlib backend (
print(matplotlib.get_backend())): pylab inline - Python version: 3.9