用逗号将一列中的单元格拆分为熊猫中的多行

Question

For an input data as follows, I want to split column office_number by comma into multiple rows:

df = pd.DataFrame({'id':['1010084420','1010084420','1010084420','1010084421','1010084421','1010084421','1010084425'],
                   'building_name': ['A', 'A', 'A', 'East Tower', 'East Tower', 'West Tower', 'T1'],
                   'floor': ['1', '1', '2', '10', '10', '11','11'],
                   'office_number':['101-105', '106', '201-203, 205, 208', '1001-1005', '1006, 1008, 1010', '1101-1103', '1101-1105'],
                   'company_name': ['Ariel Resources Ltd.', 'A.O. Tatneft', '', 'Agrium Inc.', 'Creo Products Inc.', 'Cott Corp.', 'Creo Products Inc.']})

This is my solution with reference from here :

res = (df.set_index(['id', 'building_name', 'floor', 'company_name'])
   .stack()
   .str.split(',', expand=True)
   .stack()
   .unstack(-2)
   .reset_index(-1, drop=True)
   .reset_index())

result = res[['id', 'building_name', 'floor', 'office_number', 'company_name']]

print(result)

Output:

            id building_name floor office_number          company_name
0   1010084420             A     1           106          A.O. Tatneft
1   1010084420             A     1       101-105  Ariel Resources Ltd.
2   1010084420             A     2       201-203                      
3   1010084420             A     2           205                      
4   1010084420             A     2           208                      
5   1010084421    East Tower    10     1001-1005           Agrium Inc.
6   1010084421    East Tower    10          1006    Creo Products Inc.
7   1010084421    East Tower    10          1008    Creo Products Inc.
8   1010084421    East Tower    10          1010    Creo Products Inc.
9   1010084421    West Tower    11     1101-1103            Cott Corp.
10  1010084425            T1    11     1101-1105    Creo Products Inc.

If you have any other solutions, welcome to share. Thanks.

Answer 1

Another solution is extract column by DataFrame.pop , split , stack for Series and DataFrame.join to original:

s = (df.pop('office_number')
       .str.split(',', expand=True)
       .stack()
       .reset_index(1, drop=True)
       .rename('office_number'))

res = df.join(s).reset_index(drop=True)
result = res[['id', 'building_name', 'floor', 'office_number', 'company_name']]

print(result)
            id building_name floor office_number          company_name
0   1010084420             A     1       101-105  Ariel Resources Ltd.
1   1010084420             A     1           106          A.O. Tatneft
2   1010084420             A     2       201-203                      
3   1010084420             A     2           205                      
4   1010084420             A     2           208                      
5   1010084421    East Tower    10     1001-1005           Agrium Inc.
6   1010084421    East Tower    10          1006    Creo Products Inc.
7   1010084421    East Tower    10          1008    Creo Products Inc.
8   1010084421    East Tower    10          1010    Creo Products Inc.
9   1010084421    West Tower    11     1101-1103            Cott Corp.
10  1010084425            T1    11     1101-1105    Creo Products Inc.