对我来说效果很好：

import pandas as pd
import numpy as np

data1 = np.random.randint(low=0, high=19, size=(2, 3))
data2 = np.random.randint(low=0, high=19, size=(2, 3))

df1 = pd.DataFrame(columns=['a','b','c'], data=data1)
df2 = pd.DataFrame(columns=['a','c','b'], data=data2)

column_order = ['a','b','c']
df1= df1[column_order]
df2 = df2[column_order]
df = pd.concat([df1, df2], axis=0, ignore_index=True)
df

    a   b   c
0   9   4   15
1   0   17  16
0   17  9   8
1   0   8   10

但是，如果列名不匹配，则会得到NaN值：

import pandas as pd
import numpy as np

np.random.seed(10)
data1 = np.random.randint(low=0, high=19, size=(2, 3))
data2 = np.random.randint(low=0, high=19, size=(2, 3))

df1 = pd.DataFrame(columns=['a','b','c2'], data=data1)
df2 = pd.DataFrame(columns=['a','c','b'], data=data2)

column_order = ['a','b','c']
# df1= df1[column_order]
# df2 = df2[column_order]
df = pd.concat([df1, df2], axis=0)
df

    a   b   c2       c
0   9   4   15.0    NaN
1   0   17  16.0    NaN
0   17  9   NaN 8.0
1   0   8   NaN 10.0