如何根据条件获取特定行？

Question

数据如下：

Name      Stage           Start                 End

Hulk        1      21/10/2018 06:34:15    21/10/2018 07:34:15
Hulk        2      21/10/2018 07:34:15    21/10/2018 07:54:15
Hulk        3      21/10/2018 07:58:15    21/10/2018 08:14:15
Hulk        4      21/10/2018 08:14:15    21/10/2018 08:34:15
Sam         A1     21/10/2018 09:34:15    21/10/2018 10:34:15
Sam         A2     21/10/2018 10:34:15    21/10/2018 10:45:15
Sam         A3     21/10/2018 10:45:15    21/10/2018 11:00:15
Sam         A4     21/10/2018 11:00:15    21/10/2018 11:34:15
Bruce       1.1    21/10/2018 11:34:15    21/10/2018 11:45:15
Bruce       1.2    21/10/2018 11:45:15    21/10/2018 12:00:15
Bruce       1.3    21/10/2018 12:00:15    21/10/2018 12:25:15
Bruce       1.4    21/10/2018 12:25:15    21/10/2018 12:45:15
Peter        1     21/10/2018 12:45:15    21/10/2018 01:05:15
Peter        1     21/10/2018 01:05:15    21/10/2018 01:15:15

我怎样才能为每个Name 拥有first 和last 的Stage 实例，其中以1 开头并以4 持续？

数据框应采用以下方式：

Name      Stage           Start                 End

Hulk        1      21/10/2018 06:34:15    21/10/2018 07:34:15
Hulk        4      21/10/2018 08:14:15    21/10/2018 08:34:15
Sam         A1     21/10/2018 09:34:15    21/10/2018 10:34:15
Sam         A4     21/10/2018 11:00:15    21/10/2018 11:34:15
Bruce       1.1    21/10/2018 11:34:15    21/10/2018 11:45:15
Bruce       1.4    21/10/2018 12:25:15    21/10/2018 12:45:15

我尝试了groupby 和([Name,Stage])，但没有得到上述所需的数据帧。

Answer 1

使用duplicated 和str.contains 和boolean indexing 先返回必要的行，然后使用value_counts 和map 过滤2 个行组：

m1 = ~df['Name'].duplicated()
m2 = df['Stage'].str.contains('1')

m3 = ~df['Name'].duplicated(keep='last')
m4 = df['Stage'].str.contains('4')

df1 = df[(m1 & m2) | (m3 & m4)].copy()

df1 = df1[df1['Name'].map(df1['Name'].value_counts()) == 2]
print (df1)
     Name Stage                Start                  End
0    Hulk     1  21/10/2018 06:34:15  21/10/2018 07:34:15
3    Hulk     4  21/10/2018 08:14:15  21/10/2018 08:34:15
4     Sam    A1  21/10/2018 09:34:15  21/10/2018 10:34:15
7     Sam    A4  21/10/2018 11:00:15  21/10/2018 11:34:15
8   Bruce   1.1  21/10/2018 11:34:15  21/10/2018 11:45:15
11  Bruce   1.4  21/10/2018 12:25:15  21/10/2018 12:45:15