[Bug]: Inner border is not rendered correctly when using log-scale and polar projection. #30179

Activity

1. jklymak commented on Jun 15, 2025

   

   jklymak

   on Jun 15, 2025

   Member

   More actions

   You have to call ax.set_rmin() as well
2. tillboehringer commented on Jun 15, 2025

   

   tillboehringer

   on Jun 15, 2025

   Author

   More actions

   angles = np.radians(np.linspace(15, 145, 10))
   r = np.logspace(-30, -28, len(angles))
   
   fig = plt.figure()
   ax = plt.axes(projection='polar')
   
   ax.scatter(angles, r)
   
   ax.set_theta_zero_location('N')
   ax.set_theta_direction(-1)
   ax.set_rscale('log')
   ax.set_thetamin(15)
   ax.set_thetamax(145)
   
   ax.set_rmin(10**-33)
   ax.set_rorigin(10**-33)
   

   

   angles = np.radians(np.linspace(15, 145, 10))
   r = np.logspace(-30, -28, len(angles))
   
   fig = plt.figure()
   ax = plt.axes(projection='polar')
   
   ax.scatter(angles, r)
   
   ax.set_theta_zero_location('N')
   ax.set_theta_direction(-1)
   ax.set_rscale('log')
   ax.set_thetamin(15)
   ax.set_thetamax(145)
   
   ax.set_rmin(10**-30)
   ax.set_rorigin(10**-33)
   

   

   ax.set_rmin() won't give me the correct result, also the same if I switch the order of rmin and rorigin.
3. jklymak commented on Jun 15, 2025

   

   jklymak

   on Jun 15, 2025

   Member

   More actions

   You are not being clear on what is "correct" here. The above looks correct to me.
4. rcomer commented on Jun 15, 2025

   

   rcomer

   on Jun 15, 2025

   Last edited by rcomer

   Member

   More actions

   @jklymak I think they mean the inner spine is in the wrong place - it’s all bunched up at the origin but should be at rmin.
5. tillboehringer commented on Jun 15, 2025

   

   tillboehringer

   on Jun 15, 2025

   Author

   More actions

   @jklymak yes maybe I was not clear enough. I still want there to be a radial offset, so there is an inner border like in the plot under "Expected outcome" in the initial issue statement.

   I attached a plot with the marked border, that is missing in the other plots, in red.

   
6. jklymak commented on Jun 16, 2025

   

   jklymak

   on Jun 16, 2025

   Member

   More actions

   I see. Thanks - that does seem to be a bug - not sure what is causing that.
7. dstansby

   added

   status: confirmed bug

   topic: polar

   on Jun 17, 2025
8. rcomer commented on Jun 19, 2025

   

   rcomer

   on Jun 19, 2025

   Member

   More actions

   Possibly the logic here needs to account for the axis scale

   matplotlib/lib/matplotlib/spines.py

   Lines 273 to 275 in 8b82729

   	scaled_diameter = (rmin - rorigin) / (rmax - rorigin)
   	self._height = scaled_diameter
   	self._width = scaled_diameter
9. QuLogic commented on Jun 20, 2025

   

   QuLogic

   on Jun 20, 2025

   Member

   More actions

   If you set the colour of the Axes, you can see that the Axes patch is also too big:

   import numpy as np
   import matplotlib.pyplot as plt
   
   angles = np.radians(np.linspace(15, 145, 10))
   r = np.logspace(0, 2, len(angles))
   
   fig = plt.figure()
   ax = plt.axes(projection='polar', facecolor='red')
   
   ax.scatter(angles, r)
   
   ax.set_rscale('log')
   ax.set_rorigin(10**-3)
   
   plt.show()
   

   

   The inner spine location is actually set by the Axes wedge, so the place to fix the problem is in there somewhere.
10. timhoffm

    closed this as completedin #30223on Jun 30, 2025
11. QuLogic

    added this to the v3.10.4 milestone on Jun 30, 2025