从字母表中的所有字母中选择不同的名称

Question

背景

我有以下示例 df：

df = pd.DataFrame({'Name' : ['ANT J DOE', 'ANT J DOE', 'ANT J DOE', 'ANT J DOE',
                             'ANDREW THE', 'AMANDA TO', 'AZARA HEBREW', 'BARNY GUM',
                             'BORIS CHE', 'BORIS CHE', 'BORIS CHE', 'BRYAN WILLI',
                             'BRAIN THEY', 'BREZ TOW', 'CRAFT CHEEZE', 'ERICE BEE',
                             'ERICK BEE', 'ERIC BRAI', 'ERIK REY', 'FER AZ',
                             'FRAY AR', 'HE SAM', 'JANE ANN DOE', 'JANE ANN DOE',
                             'JAMES J DOE', 'MART DR', 'QUE BALL', 'ZIM THE',
                             'ZIMMY SALLOW', 'ZAQ CRO'],
                   'P_ID': [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,
                            20,21,22,23,24,25,26,27,28,29,30]
                     })

输出

        Name     P_ID
0   ANT J DOE      1
1   ANT J DOE      2
2   ANT J DOE      3
3   ANT J DOE      4
4   ANDREW THE     5
5   AMANDA TO      6
6   AZARA HEBREW   7
7   BARNY GUM      8
8   BORIS CHE      9
9   BORIS CHE     10
10  BORIS CHE     11
11  BRYAN WILLI   12
12  BRAIN THEY    13
13  BREZ TOW      14
14  CRAFT CHEEZE  15
15  ERICE BEE     16
16  ERICK BEE     17
17  ERIC BRAI     18
18  ERIK REY      19
19  FER AZ        20
20  FRAY AR       21
21  HE SAM        22
22  JANE ANN DOE  23
23  JANE ANN DOE  24
24  JAMES J DOE   25
25  MART DR       26
26  QUE BALL      27
27  ZIM THE       28
28  ZIMMY SALLOW  29
29  ZAQ CRO       30

目标

在Name 中从每个“字母组”A-Z 中选择两个不同的名称。

例子：

1) “字母组”“A”中的两个不同名称将对应于ANT J DOE 和ANDREW THE。

2) “字母组”“Z”中的两个不同名称将对应于ZIM THE 和ZIMMY SALLOW

重要的是，Name 在此示例中并不详尽。我真正的Name 列有所有26 表示的字母表

期望的输出

        Name     P_ID
0   ANT J DOE      1
1   ANDREW THE     5
2   BARNY GUM      8
3   BORIS CHE     11
4   ERICE BEE     16
5   ERIC BRAI     18
6   FER AZ        20
7   FRAY AR       21
8   JANE ANN DOE  23
9   JAMES J DOE   25
10  ZIM THE       28
11  ZIMMY SALLOW  29

问题

如何实现我想要的输出？

Answer 1

这有点老套，我敢肯定还有更好的选择，但它确实有效：

> df['first'] = df['Name'].astype(str).str[0]  # add a column with first letter
> print( df.drop_duplicates('Name'). \         # select distinct names
    groupby('first'). \                        # group by first letter
    filter(lambda x: len(x) > 1). \            # with at least 2 entries
    groupby('first').head(2)[['Name', 'P_ID']] # select Name and P_ID columns
    )


            Name  P_ID
0      ANT J DOE     1
4     ANDREW THE     5
7      BARNY GUM     8
8      BORIS CHE     9
15     ERICE BEE    16
16     ERICK BEE    17
19        FER AZ    20
20       FRAY AR    21
22  JANE ANN DOE    23
24   JAMES J DOE    25
27       ZIM THE    28
28  ZIMMY SALLOW    29