在 matplotlib 中共享轴时显示刻度标签

Question

I'm running the following function:

def plot_variance_analysis(indices, stat_frames, legend_labels, shape):
    x = np.linspace(1, 5, 500)
    fig, axes = plt.subplots(shape[0], shape[1], sharex=True sharey=True)
    questions_and_axes = zip(indices, axes.ravel())
    frames_and_labels = zip(stat_frames, legend_labels)
    for qa in questions_and_axes:
        q = qa[0]
        ax = qa[1]
        for fl in frames_and_labels:
            frame = fl[0]
            label = fl[1]
            ax.plot(x, stats.norm.pdf(x, frame['mean'][q], frame['std'][q]), label=label)
            ax.set_xlabel(q)
            ax.legend(loc='best')
    plt.xticks([1,2,3,4,5])
    return fig, axes

Here's what I get with some of my own sample data:

I'm trying to maintain the shared state between axes, but at the same time display the tick labels for the x axis on all subplots (including the top two). I can't find any means to turn this off in the documentation. Any suggestions? Or should I just set the x tick labels axis by axis?

I'm running matplotlib 1.4.0, if that's important.

Answer 1

在 Matplotlib 2.2 及更高版本中，可以使用以下方法重新打开刻度标签：

ax.xaxis.set_tick_params(labelbottom=True)

Answer 2

The ticks that are missing have had their visible property set to False . This is pointed out in the documentation for plt.subplot . The simplest way to fix this is probably to do:

for ax in axes.flatten():
    for tk in ax.get_yticklabels():
        tk.set_visible(True)
    for tk in ax.get_xticklabels():
        tk.set_visible(True)

Here I've looped over all axes, which you don't necessarily need to do, but the code is simpler this way. You could also do this with list comprehensions in an ugly one liner if you like:

[([tk.set_visible(True) for tk in ax.get_yticklabels()], [tk.set_visible(True) for tk in ax.get_yticklabels()]) for ax in axes.flatten()]

Answer 3

You can find extra information about labels of matplotlib here: https://matplotlib.org/3.1.3/api/_as_gen/matplotlib.axes.Axes.tick_params.html

In my case, I need to turn on all the x and y labels and this solution works:

for ax in axes.flatten():
    ax.xaxis.set_tick_params(labelbottom=True)
    ax.yaxis.set_tick_params(labelleft=True)