为什么 plt.hlines() 与 transform 一起使用时会调整 ylims，而 plt.plot() 不会？

Question

I am trying to display a horizontal line at 20% of the y-axis using plt.hlines() with y=[0.2] and the transform=ax.get_xaxis_transform() argument. When I do this the upper ylim of my plot is changed to 0.2 automatically.

I do not expect this behaviour because to my understanding the get_xaxis_transform() -transform should transform the y=[0.2] value to whatever corresponds to 20% of my y-axis. In particular because the hline is displayed at 20% but the limits are adjusted anyways.

This does not happen when I perform the same operation using plt.plot() (rightmost plot). Here, the line stays at 20% of the y-axis even when interactively moving the visible area.

I am wondering why this happens and how I can avoid that hlines() adjusts the ylims when using it with the transform parameter.

MWE:

import numpy as np
import matplotlib.pyplot as plt

y = np.random.rand(20)/100  # scale so all values should be far below 0.2
x = range(len(y))

# Scenario 0 (auto-scaled axes)
fig, (ax0, ax1, ax2) = plt.subplots(ncols=3)
ax0.plot(x, y)
ax0.set(title='No horizontal lines drawn')

# Scenario 1
ax1.hlines(y=[0.2], xmin=5, xmax=15, transform=ax1.get_xaxis_transform(), color='black')
ax1.plot(x, y)
ax1.set(title='Using hlines()')

# Scenario 2
ax2.plot(x, y)
ax2.plot([5, 15], [0.2, 0.2], transform=ax2.get_xaxis_transform(), color='black')
ax2.set(title='Using plot()')

plt.show(block=True)

Answer 1

This happens because of a bug in matplotlib. https://github.com/matplotlib/matplotlib/issues/23171

As you've discovered, using plt.plot() is a simple workaround.