Updating bar and plot subplots over loop iterations

Question

I wrote the following snippet and I am trying to make it update the plots. What I get instead is an overlapping of new plots on the old ones. I researched a bit and found I needed relim() and autoscale_view(True,True,True) on the current axis. I still cannot get the desired behaviour. Is there a way to force pyplot to delete/remove the old drawing before calling plt.draw()?

import numpy as np
import matplotlib.pyplot as plt
import time

plt.ion()
a = np.arange(10)

fig,ax = plt.subplots(2,1)
plt.show()

for i in range(100):
    b = np.arange(10) * np.random.randint(10)
    ax[0].bar(a,b,align='center')
    ax[0].relim()
    ax[0].autoscale_view(True,True,True)
    ax[1].plot(a,b,'r-')
    ax[1].relim()
    ax[1].autoscale_view(True,True,True)
    plt.draw()
    time.sleep(0.01)
    plt.pause(0.001)

Answer 1

Axes has a method clear() to achieve this.

for i in range(100):
    b = np.arange(10) * np.random.randint(10)

    ax[0].clear()
    ax[1].clear()

    ax[0].bar(a,b,align='center')
    # ...

Matplotlib Axes Documentation

But relim() will always adjust your dimension to the new data so you will get a static image. Instead I would use set_ylim([min, max]) to set a fix area of values.

Answer 2

Without the need to reset the axes limits or using relim, you may want to update the bar height only.

import numpy as np
import matplotlib.pyplot as plt

plt.ion()
a = np.arange(10)

fig,ax = plt.subplots(2,1)
plt.show()

b = 10 * np.random.randint(0,10,size=10)
rects = ax[0].bar(a,b, align='center')
line, = ax[1].plot(a,b,'r-')
ax[0].set_ylim(0,100)
ax[1].set_ylim(0,100)

for i in range(100):
    b = 10 * np.random.randint(0,10,size=10)
    for rect, h in zip(rects, b):
        rect.set_height(h)
    line.set_data(a,b)
    plt.draw()
    plt.pause(0.02)