pandas，如何合并具有多列ID的两个数据帧

Question

对不起，我是熊猫新手，我很挣扎。基本上，我有两个数据集df1:

站点	nb样本
A	3
B	2
C	1

使用pandas.DataFrame.merge和.rename()的单行

df_new = df2.merge(df1, left_on='sites1', right_on='sites', how='left').merge(df1, left_on='sites2', right_on='sites', how='left')[['sites1', 'sites2', 'nb links', 'nb samples_x', 'nb samples_y']].rename(columns={'nb samples_x': 'nb samples sites1', 'nb samples_y': 'nb samples sites2'})
[Out]:
sites1 sites2  nb links  nb samples sites1  nb samples sites2
0      A      B         3                  3                  2
1      A      C         1                  3                  1

---

票据

• 让我们分解一下，让它更容易理解：

  1. 通过合并两个数据帧开始

     df_new = df2.merge(df1, left_on='sites1', right_on='sites', how='left').merge(df1, left_on='sites2', right_on='sites', how='left')
     [Out]:
     sites1 sites2  nb links sites_x  nb samples_x sites_y  nb samples_y
     0      A      B         3       A             3       B             2
     1      A      C         1       A             3       C             1
     

  2. 只选择要考虑的列

     df_new = df_new[['sites1', 'sites2', 'nb links', 'nb samples_x', 'nb samples_y']]
     [Out]:
     sites1 sites2  nb links  nb samples_x  nb samples_y
     0      A      B         3             3             2
     1      A      C         1             3             1
     

  3. 重命名列

     df_new.columns = ['sites1', 'sites2', 'nb links', 'nb samples sites1', 'nb samples sites2']
     [Out]:
     sites1 sites2  nb links  nb samples sites1  nb samples sites2
     0      A      B         3                  3                  2
     1      A      C         1                  3                  1