熊猫合并不保持"开"栏

Question

我正在尝试将两个数据帧合并pandas到一个公共列名称(orderid)上.结果数据帧(合并的数据帧)正在从第二个数据帧中删除orderid.根据文档,除非您明确告知不要,否则应保留"on"列.

import pandas as pd    
df = pd.DataFrame([[1,'a'], [2, 'b'], [3, 'c']], columns=['orderid', 'ordervalue'])
df['orderid'] = df['orderid'].astype(str)
df2 = pd.DataFrame([[1,200], [2, 300], [3, 400], [4,500]], columns=['orderid', 'ordervalue'])
df2['orderid'] = df2['orderid'].astype(str)
pd.merge(df, df2, on='orderid', how='outer', copy=True, suffixes=('_left', '_right'))

哪个输出:

|      |orderid | ordervalue_left | ordervalue_right |
|------|--------|-----------------|------------------|
| 0    | 1      | a               | 200              |
| 1    | 2      | b               | 300              |
| 2    | 3      | c               | 400              |
| 3    | 4      |                 | 500              |

我想要创建的是:

|      | orderid_left | ordervalue_left | orderid_left | ordervalue_right |
|------|--------------|-----------------|--------------|------------------|
| 0    | 1            | a               | 1            | 200              |
| 1    | 2            | b               | 2            | 300              |
| 2    | 3            | c               | 3            | 400              |
| 3    | NaN          | NaN             | 4            | 500              |

我该怎么写呢？

Answer 1

重命名这些orderid列，以便df有一个名为 的列orderid_left，并df2有一个名为 的列orderid_right：

import pandas as pd    
df = pd.DataFrame([[1,'a'], [2, 'b'], [3, 'c']], columns=['orderid', 'ordervalue'])
df['orderid'] = df['orderid'].astype(str)
df2 = pd.DataFrame([[1,200], [2, 300], [3, 400], [4,500]], columns=['orderid', 'ordervalue'])
df2['orderid'] = df2['orderid'].astype(str)

df = df.rename(columns={'orderid':'orderid_left'})
df2 = df2.rename(columns={'orderid':'orderid_right'})
result = pd.merge(df, df2, left_on='orderid_left', right_on='orderid_right', 
                  how='outer', suffixes=('_left', '_right'))
print(result)

产量

  orderid_left ordervalue_left orderid_right  ordervalue_right
0            1               a             1               200
1            2               b             2               300
2            3               c             3               400
3          NaN             NaN             4               500