分层索引按列排序,同时保留原始列

我正试图模拟一个Hirarical索引数据帧,如下所示:

>>> raw_data = ({'city': ['Delhi', 'Kanpur', 'Mumbai', 'Pune','Delhi', 'Kanpur', 'Mumbai', 'Pune'],
...                 'rank': ['1st', '2nd', '1st', '2nd','1st', '2nd', '1st', '2nd'],
...                 'name': ['Ramesh', 'Kirpal', 'Jungi', 'Sanju','Ramesh', 'Kirpal', 'Jungi', 'Sanju'],
...                 'score1': [10,15,20,25,10,15,20,25],
...                 'score2': [20,35,40,45,20,35,40,45]})

下面是DataFrame的样子,所以datFrame附带了默认索引。

>>> df = pd.DataFrame(raw_data, columns = ['city', 'rank', 'name', 'score1', 'score2'])
>>> df
     city rank    name  score1  score2
0   Delhi  1st  Ramesh      10      20
1  Kanpur  2nd  Kirpal      15      35
2  Mumbai  1st   Jungi      20      40
3    Pune  2nd   Sanju      25      45
4   Delhi  1st  Ramesh      10      20
5  Kanpur  2nd  Kirpal      15      35
6  Mumbai  1st   Jungi      20      40
7    Pune  2nd   Sanju      25      45

我想通过选择 'city', 'rank' set.index 方法,同时保留原始列。

>>> df.set_index(['city', 'rank'], drop=False)
               city rank    name  score1  score2
city   rank
Delhi  1st    Delhi  1st  Ramesh      10      20
Kanpur 2nd   Kanpur  2nd  Kirpal      15      35
Mumbai 1st   Mumbai  1st   Jungi      20      40
Pune   2nd     Pune  2nd   Sanju      25      45
Delhi  1st    Delhi  1st  Ramesh      10      20
Kanpur 2nd   Kanpur  2nd  Kirpal      15      35
Mumbai 1st   Mumbai  1st   Jungi      20      40
Pune   2nd     Pune  2nd   Sanju      25      45

但我希望有索引 city 首先,然后按索引 rank

               city rank    name  score1  score2
city   rank
Delhi  1st    Delhi  1st  Ramesh      10      20
       1st    Delhi  1st  Ramesh      10      20

Kanpur 2nd   Kanpur  2nd  Kirpal      15      35
       2nd   Kanpur  2nd  Kirpal      15      35

Mumbai 1st   Mumbai  1st   Jungi      20      40
       1st   Mumbai  1st   Jungi      20      40

Pune   2nd     Pune  2nd   Sanju      25      45
       2nd     Pune  2nd   Sanju      25      45