没有可传递给df.plot的参数，该参数对单个列的柱形着色不同.
由于不同列的条的颜色不同，因此一种选择是在绘制之前转置数据框，

There is no argument you can pass to df.plot that colorizes the bars differently for a single column.
Since bars for different columns are colorized differently, an option is to transpose the dataframe before plotting,

ax = df.T.plot(kind='bar', label='index', colormap='Paired')

现在这会将数据绘制为子组的一部分.因此，需要进行一些调整才能正确设置限制和xlabel.

This would now draw the data as part of a subgroup. Therefore some tweaking needs to be applied to set the limits and xlabels correctly.

import matplotlib.pyplot as plt
import pandas as pd

df = pd.DataFrame({'count': {0: 3372, 1: 68855, 2: 17948, 3: 708, 4: 9117}}).reset_index()

ax = df.T.plot(kind='bar', label='index', colormap='Paired')
ax.set_xlim(0.5, 1.5)
ax.set_xticks([0.8,0.9,1,1.1,1.2])
ax.set_xticklabels(range(len(df)))
plt.show()

虽然我猜这个解决方案符合问题的标准，但是使用plt.bar实际上并没有错.只需呼叫plt.bar就足够了

While I guess this solution matches the criteria from the question, there is actually nothing wrong with using plt.bar. A single call to plt.bar is sufficient

plt.bar(range(len(df)), df["count"], color=plt.cm.Paired(np.arange(len(df))))

完整代码:

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

df = pd.DataFrame({'count': {0: 3372, 1: 68855, 2: 17948, 3: 708, 4: 9117}}).reset_index()

plt.bar(range(len(df)), df["count"], color=plt.cm.Paired(np.arange(len(df))))

plt.show()