python中pandas对多列进行分组统计的实现

import pandas as pd
res6 = pd.read_csv('test.csv')
res6.shape

(12, 3)

res6.columns

Index(['user_id', 'cate', 'shop_id'], dtype='object')

res6.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 12 entries, 0 to 11
Data columns (total 3 columns):
user_id    12 non-null int64
cate       12 non-null int64
shop_id    12 non-null int64
dtypes: int64(3)
memory usage: 368.0 bytes

res6.describe()

res6

res6['user_id'].value_counts()

390192     3
982499     2
493873     2
242141     1
1021819    1
533696     1
1558165    1
283284     1
Name: user_id, dtype: int64

res6.groupby(['user_id']).size().sort_values(ascending=False)

user_id
390192     3
982499     2
493873     2
1558165    1
1021819    1
533696     1
283284     1
242141     1
dtype: int64

res6.groupby(['user_id', 'cate']).size().sort_values(ascending=False)

user_id  cate
982499   7       2
493873   7       2
1558165  7       1
1021819  7       1
533696   7       1
390192   23      1
         22      1
         20      1
283284   7       1
242141   7       1
dtype: int64

res6_test = res6.groupby(['user_id', 'cate', 'shop_id']).size().sort_values(ascending=False)
res6_test

user_id  cate  shop_id
493873   7     3388       2
1558165  7     3388       1
1021819  7     3388       1
982499   7     3389       1
               3388       1
533696   7     3388       1
390192   23    4179       1
         22    4278       1
         20    4178       1
283284   7     3388       1
242141   7     3388       1
dtype: int64