Open
Description
Hi! The following code breaks:
import seaborn as sns
import matplotlib.pyplot as plt
def create_point_plot():
df = sns.load_dataset("anscombe")
# simulate a dataset with n_hue_levels == 1
df = df[df["dataset"] == "I"]
sns.pointplot(data=df, x="x", y="y", hue="dataset", dodge=True)
plt.show()
create_point_plot()This is the error message:
Traceback (most recent call last):
File "/home/cdp58/Documents/repos/pasna_analysis/pasna_analysis/seaborn_zeroIndex.py", line 14, in <module>
create_point_plot()
File "/home/cdp58/Documents/repos/pasna_analysis/pasna_analysis/seaborn_zeroIndex.py", line 10, in create_point_plot
sns.pointplot(data=df, x="x", y="y", hue="dataset", dodge=True)
File "/home/cdp58/miniconda3/envs/pscope_analysis/lib/python3.11/site-packages/seaborn/categorical.py", line 2516, in pointplot
p.plot_points(
File "/home/cdp58/miniconda3/envs/pscope_analysis/lib/python3.11/site-packages/seaborn/categorical.py", line 1218, in plot_points
step_size = dodge / (n_hue_levels - 1)
~~~~~~^~~~~~~~~~~~~~~~~~~~
ZeroDivisionError: float division by zero
While it does not make strict sense to have dodge=True and a single categorical variable, it would be nice to be able to handle this case. At least this happens sometimes with the datasets I have... I usually have multiple groups, but sometimes only a single group. I understand this could be handled on my end.
I believe flipping dodge to False inside plot_points if n_hue_levels <= 1 would be enough.
I ran this using seaborn=0.13.2
Metadata
Metadata
Assignees
Labels
No labels