Question

如何使用多索引创建具有聚合值的新列？

例如，在以下DataFrame中，我如何使用基于区域，市场的索引的聚合产品列表创建新列？

import pandas as pd

df = pd.DataFrame({'product' : ['Alpha', 'Alpha', 'Beta', 'Beta', 'Omega', 'Omega', 'Delta', 'Delta'],
               'region' : [1, 2, 1, 1, 3, 1, 2, 1],
               'market' : ['small', 'large', 'small', 'small', 'large', 'small', 'small', 'medium']})

希望从：


    +---------+--------+--------+
    | product | region | market |
    +---------+--------+--------+
    | Alpha   |      1 | small  |
    | Alpha   |      2 | large  |
    | Beta    |      1 | small  |
    | Beta    |      1 | small  |
    | Omega   |      3 | large  |
    | Omega   |      1 | small  |
    | Delta   |      2 | small  |
    | Delta   |      1 | medium |
    +---------+--------+--------+

要：


    +---------+--------+--------+----------------------------+
    | product | region | market |          product_list      |
    +---------+--------+--------+----------------------------+
    | Alpha   |      1 | small  | ['Alpha', 'Beta', 'Omega'] |
    | Alpha   |      2 | large  | ['Alpha']                  |
    | Beta    |      1 | small  | ['Alpha', 'Beta', 'Omega'] |
    | Beta    |      1 | small  | ['Alpha', 'Beta', 'Omega'] |
    | Omega   |      3 | large  | ['Omega']                  |
    | Omega   |      1 | small  | ['Alpha', 'Beta', 'Omega'] |
    | Delta   |      2 | small  | ['Delta']                  |
    | Delta   |      1 | medium | ['Delta']                  |
    +---------+--------+--------+----------------------------+

重复丢弃（例如Beta，1，小而不是在product_list中复制[＆＃39; Alpha＆＃39;，＆＃39; Beta＆＃39;，＆＃39; Omega＆＃39;] < / p>

Answer 1

IIUC您可以将groupby与transform一起使用，通过转换为set删除重复项，然后使用apply转换为list：

df['product_list'] = df.groupby(['region','market']).transform(lambda x:
                                                                set(x.tolist()))
df['product_list'] = df['product_list'].apply(lambda x: list(x))
print df

   market product  region          product_list
0   small   Alpha       1  [Alpha, Beta, Omega]
1   large   Alpha       2               [Alpha]
2   small    Beta       1  [Alpha, Beta, Omega]
3   small    Beta       1  [Alpha, Beta, Omega]
4   large   Omega       3               [Omega]
5   small   Omega       1  [Alpha, Beta, Omega]
6   small   Delta       2               [Delta]
7  medium   Delta       1               [Delta]

Pandas多索引查找

1 个答案: