合并大熊猫中的数据帧

时间:2014-05-14 13:48:25

标签: python pandas merge

pandas.merge对左右两侧的行为有所不同!对于左侧,如果我们一起使用left_on和left_index它会显示错误,但右侧的工作相同!!!

代码:

import pandas as pd
import numpy as np
right = pd.DataFrame(data=np.arange(12).reshape((6,2)),index=[['Nevada', 'Nevada', 'Ohio', 'Ohio', 'Ohio', 'Ohio'],[2001, 2000, 2000, 2000, 2001, 2002]],columns=['event1','event2'])
left = pd.DataFrame(data={'key1':['Ohio', 'Ohio', 'Ohio', 'Nevada', 'Nevada'],'key2':[2000, 2001, 2002, 2001, 2002],'data':np.arange(5.)})
pd.merge(left,right,right_index=True,left_index=True,right_on='event1')#it works and returns an empty table which is expected
pd.merge(left,right,left_index=True,right_index=True,left_on='key1')# it makes error !!!

1 个答案:

答案 0 :(得分:1)

你有几个问题正在发生。首先,您的合并语句未正确构造。您不应同时使用left_onleft_indexright_onright_index。您应该只使用一个左选项和一个右选项。

您在第二个语句中出错的原因是索引级别不匹配。在左侧合并中,左侧索引是单个级别,当您同时指定right_index=Trueright_on='event1'时,right_on属性优先。由于两者都是单级整数,所以没有问题。我应该指出合并,如果构造正确,(pd.merge(left, right, left_index=True, right_on='event1', how='left'))不会产生空的DataFrame ...请参阅下面的代码。

在右侧合并中,您使用right_index=True指定正确的索引,left_on优先于left_index=True。这里的问题是正确的索引是2级,其中' key1`字段是单级字符串。

In [1]: import pandas as pd

In [2]: import numpy as np

In [3]: right = pd.DataFrame(data=np.arange(12).reshape((6,2)),index=[['Nevada', 'Nevada', 'Ohio', 'Ohio', 'Ohio', 'Ohio'],[2001, 2000, 2000, 2000, 2001, 2002]],columns=['event1','event2'])

In [4]: left = pd.DataFrame(data={'key1':['Ohio', 'Ohio', 'Ohio', 'Nevada', 'Nevada'],'key2':[2000, 2001, 2002, 2001, 2002],'data':np.arange(5.)})

In [5]: left
Out[5]:
   data    key1  key2
0     0    Ohio  2000
1     1    Ohio  2001
2     2    Ohio  2002
3     3  Nevada  2001
4     4  Nevada  2002

In [6]: right
Out[6]:
             event1  event2
Nevada 2001       0       1
       2000       2       3
Ohio   2000       4       5
       2000       6       7
       2001       8       9
       2002      10      11

In [5]: left_merge = left.merge(right, left_index=True, right_on='event1', how='left')

In [7]: left_merge
Out[7]:
             data    key1  key2  event1  event2
Nevada 2001     0    Ohio  2000       0       1
Ohio   2002     1    Ohio  2001       1     NaN
Nevada 2000     2    Ohio  2002       2       3
Ohio   2002     3  Nevada  2001       3     NaN
       2000     4  Nevada  2002       4       5