Broken axis slash marks inside bar chart in matplotlib?

问题

I've seen matplotlib examples of placing the broken axis slash marks on the axes, such as this one.

My questions, how can I place it where the bars are broken? Can this be done in a programmatic way for updating time-series plots month over month?

Below is an example of what I want done, using excel. Notice the tildes in the Paris June bar and the Madrid May bar. The tildes hide a portion of the bar it's covering.

broken axis excel example

I'm also providing easier sample data and what I've been able to do so far.

XX = pd.Series([200,400,100,1400],index=['x1','x2','x3','x4'])
fig, (ax1,ax2) = plt.subplots(2,1,sharex=True,
                         figsize=(5,6))
ax1.spines['bottom'].set_visible(False)
ax1.tick_params(axis='x',which='both',bottom=False)
ax2.spines['top'].set_visible(False)
ax2.set_ylim(0,500)
ax1.set_ylim(1200,1500)
ax1.set_yticks(np.arange(1000,1501,100))
XX.plot(ax=ax1,kind='bar')
XX.plot(ax=ax2,kind='bar')
for tick in ax2.get_xticklabels():
    tick.set_rotation(0)
d = .015  
kwargs = dict(transform=ax1.transAxes, color='k', clip_on=False)
ax1.plot((-d, +d), (-d, +d), **kwargs)      
ax1.plot((1 - d, 1 + d), (-d, +d), **kwargs)
kwargs.update(transform=ax2.transAxes)  
ax2.plot((-d, +d), (1 - d, 1 + d), **kwargs)  
ax2.plot((1 - d, 1 + d), (1 - d, 1 + d), **kwargs)
plt.show()

回答1:

Just to show the principle, you can put the same kind of line at every position where a bar exceeds the upper limit of the lower axes and also where a bar exceeds the lower limit of the upper axes.

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np

XX = pd.Series([200,400,100,1400],index=['x1','x2','x3','x4'])
fig, (ax1,ax2) = plt.subplots(2,1,sharex=True,
                         figsize=(5,6))
ax1.spines['bottom'].set_visible(False)
ax1.tick_params(axis='x',which='both',bottom=False)
ax2.spines['top'].set_visible(False)

bs = 500
ts = 1000

ax2.set_ylim(0,bs)
ax1.set_ylim(ts,1500)
ax1.set_yticks(np.arange(1000,1501,100))

bars1 = ax1.bar(XX.index, XX.values)
bars2 = ax2.bar(XX.index, XX.values)

for tick in ax2.get_xticklabels():
    tick.set_rotation(0)
d = .015  
kwargs = dict(transform=ax1.transAxes, color='k', clip_on=False)
ax1.plot((-d, +d), (-d, +d), **kwargs)      
ax1.plot((1 - d, 1 + d), (-d, +d), **kwargs)
kwargs.update(transform=ax2.transAxes)  
ax2.plot((-d, +d), (1 - d, 1 + d), **kwargs)  
ax2.plot((1 - d, 1 + d), (1 - d, 1 + d), **kwargs)

for b1, b2 in zip(bars1, bars2):
    posx = b2.get_x() + b2.get_width()/2.
    if b2.get_height() > bs:
        ax2.plot((posx-3*d, posx+3*d), (1 - d, 1 + d), color='k', clip_on=False,
                 transform=ax2.get_xaxis_transform())
    if b1.get_height() > ts:
        ax1.plot((posx-3*d, posx+3*d), (- d, + d), color='k', clip_on=False,
                 transform=ax1.get_xaxis_transform())
plt.show()

It doesn't look great, but can of course be adapted with a nicer shape.

来源：https://stackoverflow.com/questions/53642861/broken-axis-slash-marks-inside-bar-chart-in-matplotlib

标签

python

matplotlib

bar-chart