【翻译】：UserWarning:FixedFormatter should only be used together with FixedLocator

问题分析

1）问题释义：问题提示的意思就是，指定label格式 即FixedFormatter的时候，需要同时指定这些label的位置/FixedLocator。
2）解决方案：简单来说，用set_ticks之后就不会出现该提示；但是它会改变原有的tick&label的位置，这是就用到了FixedLocator。具体展示例子如下：

Answered by Rational-IM

import numpy as np
import matplotlib.pyplot as plt
import matplotlib.ticker as mticker

x = np.array(range(1000, 5000, 500))
y = 37*x

# 首先，列好图画基础
fig, [ax1, ax2, ax3] = plt.subplots(1, 3, figsize=(12,6))
ax1.plot(x,y, linewidth=5, color='green')
ax2.plot(x,y, linewidth=5, color='red')
ax3.plot(x,y, linewidth=5, color='blue')

# ax1不做任何修改，当一个"control chart"做对比用

# 使用"set_yticks"更新yticks。体现在ax2上
ax2.set_yticks(ax1.get_yticks())  # 和ax1使用一样的个数的yticks。 get_xticks: Return the xaxis' tick locations in data coordinates
# ax1.get_yticks() 结果[ 20000.  40000.  60000.  80000. 100000. 120000. 140000. 160000. 180000.]
ax2.set_yticklabels([f"{x:,.0f}" for x in ax2.get_yticks()])

# 使用matplotlib.ticker.FixedLocator 更新yticks。体现在ax3上
ax3.yaxis.set_major_locator(mticker.FixedLocator(ax1.get_yticks()))
ticks_loc_ax3 = ax3.get_yticks()
ax3.set_yticklabels([f"{x:,.0f}" for x in ticks_loc_ax3])
# 同时搭配使用"MaxNLocator"更新xticks(调整xtick的个数)
ax3.xaxis.set_major_locator(mticker.MaxNLocator(3))
ticks_loc_ax3 = ax3.get_xticks()
ax3.xaxis.set_major_locator(mticker.FixedLocator(ticks_loc_ax3))
ax3.set_xticklabels([f"{x:,.0f}" for x in ticks_loc_ax3])

plt.tight_layout()
plt.show()

对于共x轴的主副坐标轴来说，这样也行：

fig, ax1 = plt.subplots()
ax1.plot(<blabla...>, label='label_of_ax1')
ax1.set_xticks(<list_of_ax1_xticks>)  # 确定xticks的个数
ax1.set_xticklabels(<list_of_ax1_xticks_labels>, rotation=60) # 确定xticks都显示什么labels
plt.ylim(min_of_ax1y, max_of_ax1y)
plt.legend(loc='upper right')

ax2 = ax1.twinx()
ax2.plot(<blabla...>, label='label_of_ax2')
ax2.set_xticks(<list_of_ax2_xticks>)
plt.legend(loc=(0.85, 0.78))
plt.ylim(min_of_ax2y, max_of_ax2y)

ax1.grid(axis='x')
plt.show()

Axes.set_yticks(ticks, labels=None, *, minor=False, **kwargs)

Set the yaxis’ tick locations and optionally labels.

If necessary, the view limits of the Axis are expanded so that all given ticks are visible.