1）例如，我有 3 列，如下所示 date categories contents 2018-01 fish_tank1 Goldfish Gombessa Goosefish Gopher rockfish 2018-01 fish_tank2 Grass carp Goosefish Grayling mullet shark 2018-02 fish_tank2 Goosefish Gopher rockfish Grayling mullet shark 2018-01 fish_tank1 carp Goosefish Grayling Goldfish Gombessa 2018-02 fish_tank2 carp Goosefish Grayling Grass carp Goosefish 2018-03 fish_tank3 Grass carp Goosefish Grayling mullet shark 2018-03 fish_tank2 Goosefish Gopher rockfish Goosefish Grayling 2）我有点想做df.groupby(['date','categories']).agg(df.contents.str.split(expand=True).stack().value_counts()得到类似下面的结果。但最近几天我无法弄清楚这一点。 date categories contents 2018-01 fish_tank1 2 Goldfish 2 Gombessa 2 Goosefish 2 Gopher 1 rockfish 1 ...... fish_tank2 Grass 1 carp 1 ..... 2018-02 fish_tank2 Goosefish 3 Grayling 2 Gopher 1 ........ ........................ 3）谁能给我洞察力来获得我想做的结果？
1 回答
largeQ
TA贡献2039条经验 获得超7个赞
用 -
from collections import Counter
df['contents2'] = df['contents'].str.split()
df.groupby(['date', 'categories'])['contents2'].apply(lambda x: Counter(x.sum()))
输出
date categories
2018-01 fish_tank1 Goldfish 2.0
Gombessa 2.0
Goosefish 2.0
Gopher 1.0
Grayling 1.0
carp 1.0
rockfish 1.0
fish_tank2 Goosefish 1.0
Grass 1.0
Grayling 1.0
carp 1.0
mullet 1.0
shark 1.0
2018-02 fish_tank2 Goosefish 3.0
Gopher 1.0
Grass 1.0
Grayling 2.0
carp 2.0
mullet 1.0
rockfish 1.0
shark 1.0
2018-03 fish_tank2 Goosefish 2.0
Gopher 1.0
Grayling 1.0
rockfish 1.0
fish_tank3 Goosefish 1.0
Grass 1.0
Grayling 1.0
carp 1.0
mullet 1.0
shark 1.0
Name: contents2, dtype: float64
添加回答
举报
0/150
提交
取消