python - 在 Python 中合并 Pandas DataFrame

Question

我有两个 pandas 数据框。第一个包含 3401 行 1 列，第二个包含 4 行 3 列。

但是我得到的是(我的脚本的输出示例):

 DataFrame1      |     DataFrame2 

<小时/>

 - email1        |     -Id1  -Project1 -Descr1
 - email2        |     -Id2  -Project2 -Descr2
 - email3        |     -Id3  -Project3 -Descr3
 - email4        |     -Id4  -Project4 -Descr4
 - email5        |     -None -None     -None 
  ... ....       |      ... ...   
 - email3401     |     -None -None     -None

我想做的是对于每封邮件，我都想得到这样的东西:

 - mail1, Id1, Project1, Descr1, Id2, Project2, ... , Id4, Project4, Descr4
 - mail2, Id1, Project1, Descr1, Id2, Project2, ... , Id4, Project4, Descr4
 ... ...
 - mail3401, Id1, Project1, Descr1, Id2, Project2, ... , Id4, Project4, Descr4 

感谢您的建议!

这是我的代码:

     path = r"/Users/kd/path"
     allFiles = glob.glob(path + "/*.csv")
     frame = pd.DataFrame()
     file_names = []
     j=0
     for file_ in allFiles:
         name = os.path.splitext(file_)[0]
         i = int(name[-1])
         file_names.append(name)
         df = pd.read_csv(file_, index_col = None, header = 0)
         if j>0:
            globals()["self.dfInternautes%s"%i] =   pd.concat([globals(["self.dfInternautes%s"%i], df], axis=1)
         else: 
            globals()["self.dfInternautes%s"%i] = df
         j += 1

Answer 1

要从 DataFrame 中生成一行，请使用 stack。然后对其进行迭代，在第一个 DataFrame 中创建新列。

>>> df1
        0
0  email1
1  email2
2  email3
3  email4
4  email5
5  email6
>>> df2
     0         1       2
0  Id1  Project1  Descr1
1  Id2  Project2  Descr2
2  Id3  Project3  Descr3
3  Id4  Project4  Descr4
>>> st = df2.stack()
>>> st
0  0         Id1
   1    Project1
   2      Descr1
1  0         Id2
   1    Project2
   2      Descr2
2  0         Id3
   1    Project3
   2      Descr3
3  0         Id4
   1    Project4
   2      Descr4
dtype: object
>>> df = df1.copy()
>>> for i in st.index: df[i] = st[i]
... 
>>> df
        0 (0, 0)    (0, 1)  (0, 2) (1, 0)    (1, 1)  (1, 2) (2, 0)    (2, 1)  \
0  email1    Id1  Project1  Descr1    Id2  Project2  Descr2    Id3  Project3   
1  email2    Id1  Project1  Descr1    Id2  Project2  Descr2    Id3  Project3   
2  email3    Id1  Project1  Descr1    Id2  Project2  Descr2    Id3  Project3   
3  email4    Id1  Project1  Descr1    Id2  Project2  Descr2    Id3  Project3   
4  email5    Id1  Project1  Descr1    Id2  Project2  Descr2    Id3  Project3   
5  email6    Id1  Project1  Descr1    Id2  Project2  Descr2    Id3  Project3   

   (2, 2) (3, 0)    (3, 1)  (3, 2)  
0  Descr3    Id4  Project4  Descr4  
1  Descr3    Id4  Project4  Descr4  
2  Descr3    Id4  Project4  Descr4  
3  Descr3    Id4  Project4  Descr4  
4  Descr3    Id4  Project4  Descr4  
5  Descr3    Id4  Project4  Descr4  

可以选择更改列名称

df.columns = ['email', 'Id1', 'Project1', 'Descr1', 'Id2', 'Project2', 'Descr2', 'Id3', 'Project3', 'Descr3', 'Id4', 'Project4', 'Descr4']