Matplotlib：使用canvas.draw（）重新绘制3D图形时的附加轴

Question

I have what is probably a very simple problem replotting some 3D data using Matplotlib. Initially, I have an figure with a 3D projection on a canvas:

self.fig = plt.figure()
self.canvas = FigCanvas(self.mainPanel, -1, self.fig)
self.axes = self.fig.add_subplot(111, projection='3d')

I then add some data and use canvas.draw() to update. The plot itself updates as expected, but I get additional 2D axis on the outside of the figure (-0.05 to 0.05) and I can't work out how to stop it:

self.axes.clear()
self.axes = self.fig.add_subplot(111, projection='3d')

xs = np.random.random_sample(100)
ys = np.random.random_sample(100)
zs = np.random.random_sample(100)

self.axes.scatter(xs, ys, zs, c='r', marker='o')
self.canvas.draw()

Any ideas? I'm going in circles right now!

Answer 1

Instead of axes.clear() + fig.add_subplot , use the remove method of the mpl_toolkits.mplot3d.art3d.Patch3DCollection object:

In [31]: fig = plt.figure()

In [32]: ax = fig.add_subplot(111, projection='3d')

In [33]: xs = np.random.random_sample(100)

In [34]: ys = np.random.random_sample(100)

In [35]: zs = np.random.random_sample(100)

In [36]: a = ax.scatter(xs, ys, zs, c='r', marker='o')   #draws

In [37]: a.remove()                                      #clean

In [38]: a = ax.scatter(xs, ys, zs, c='r', marker='o')   #draws again

If you still have problems you can play with this:

import numpy as np
from matplotlib import pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
from matplotlib import interactive
interactive(True)

xs = np.random.random_sample(100)
ys = np.random.random_sample(100)
zs = np.random.random_sample(100)

fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')

a = ax.scatter(xs, ys, zs, c='r', marker='o')

plt.draw()

raw_input('press for new image')

a.remove()

xs = np.random.random_sample(1000)
ys = np.random.random_sample(1000)
zs = np.random.random_sample(1000)

a = ax.scatter(xs, ys, zs, c='r', marker='o')

plt.draw()

raw_input('press to end')

Answer 2

Joquin's suggestions worked well and highlighted that I was probably going about plotting the wrong way to start with. However, for the sake of completeness, I eventually found that you can get rid of the 2D axis simply by using:

self.axes.get_xaxis().set_visible(False)
self.axes.get_yaxis().set_visible(False)

This seems to be one way at least of removing the 2D labels from 3D plots if they appear.