Pandas concat DataFrames - 保持索引的原始顺序

时间:2017-02-21 07:55:43

标签: python pandas dataframe concat

我正在尝试连接两个数据帧:

df2:
CU  Pmt  2017-02-01
h    b        15
h    d        12
h    a        13

和df1:

CU  Pmt  'Total/Max/Min'
h    b        20
h    d        23 
h    a        22
a    b        16
a    d        13
a    a        14

这样df3:

CU  Pmt  2017-02-01  2017-02-02
h    b      15           20
h    d      12           23
h    a      13           22
a    b      NaN          16
a    d      NaN          13
a    a      Nan          14

我正在为

使用index_col = [0,1]的多索引

这就是我所拥有的:

date = '2017-02-02'
df1 = pd.read_csv(r'Data\2017-02\2017-02-02\Aggregated\Aggregated_Daily_All.csv', usecols=['CU', 'Parameters', 'Total/Max/Min'], index_col =[0,1])
df1 = df1.rename(columns = {'Total/Max/Min':date})

df2 = pd.read_csv(r'Data\2017-02\MonthlyData\February2017.csv', index_col = [0,1])
df3  = pd.concat([df2, df1], axis=1)
df3.to_csv(r'Data\2017-02\MonthlyData\February2017.csv')

然而,df3的出现是:

CU  Pmt  2017-02-01  2017-02-02
a    a      NaN          14
a    b      NaN          16
a    d      Nan          13
h    a      13           22
h    b      15           20
h    d      12           23

其中CUPmt(两个索引列)按字母顺序排列。如何保留原始订单,以便为新日期添加的所有新索引都添加到底部?

1 个答案:

答案 0 :(得分:2)

如果df1.index的值包含df2.index的值,则可以尝试reindex

df3  = pd.concat([df2, df1], axis=1).reindex(df1.index)
print (df3)
        2017-02-01  'Total/Max/Min'
CU Pmt                             
h  b          15.0               20
   d          12.0               23
   a          13.0               22
a  b           NaN               16
   d           NaN               13
   a           NaN               14