将新的Pandas DataFrame附加到旧的，而没有对列名称进行排序的情况

Question

I append a new dataframe to an old one:

import numpy as np
import pandas as pd
from pandas import Series
from pandas import DataFrame

df1 = DataFrame(np.arange(3.).reshape((1, 3)), columns=list('dcb'), index=['Ohio'])
df2 = DataFrame(np.arange(3.).reshape((1, 3)), columns=list('bdc'), index=['Utah'])
print df1
print df2
print pd.concat([df1, df2])

then i got result like this:

       d    c    b
Ohio  0.0  1.0  2.0

       b    d    c
Utah  0.0  1.0  2.0

       b    c    d
Ohio  2.0  1.0  0.0
Utah  0.0  2.0  1.0

however i want the columns in the result not sorted as 'bcd' but as origin 'dcb' like:

       d    c    b
Ohio  0.0  1.0  2.0
Utah  1.0  2.0  0.0

Answer 1

使用join_axes参数：

pd.concat([df1, df2], join_axes=[df1.columns])

Answer 2

You can store the original order in a variable and then reapply it after combining:

df1 = DataFrame(np.arange(3.).reshape((1, 3)), columns=list('dcb'), index=['Ohio'])
orig_column_order = df1.columns
df2 = DataFrame(np.arange(3.).reshape((1, 3)), columns=list('bdc'), index=['Utah'])
combined = pd.concat([df1, df2], keys=list('dbc'))
combined = combined[orig_column_order]
print(df1)
print(df2)
print(combined)

Gives:

        d    c    b
Ohio  0.0  1.0  2.0
        b    d    c
Utah  0.0  1.0  2.0
          d    c    b
d Ohio  0.0  1.0  2.0
b Utah  1.0  2.0  0.0