如何在group by之后单独获取计数

Question

df_n_gender_grp = df_n_gender_prod_cat.groupby(['Gender','prod_cat'])

我的输出如下所示

    Gender  prod_cat
0   M   Books
1   M   Books
2   M   Electronics
3   M   Electronics
4   M   Books
100 F   Electronics
101 F   Electronics
102 F   Electronics
103 F   Electronics
104 F   Electronics
105 F   Clothing
106 F   Clothing
107 F   Clothing
108 F   Clothing

我的预期输出是

按值的降序排列的第一个数据帧

M Books       2
M Electronics 3

按值降序排列的第二个数据帧

F Electronics 5
F Clothing    4

Answer 1

用于带有MultiIndex GroupBy.size和Series.sort_index的Series

s = df_n_gender_prod_cat.groupby(['Gender','prod_cat']).size().sort_index(ascending=False)
print (s)
Gender  prod_cat   
M       Electronics    2
        Books          3
F       Electronics    5
        Clothing       4
dtype: int64

然后按loc过滤

df1 = s.loc[['F']]
df2 = s.loc[['M']]

对于DataFrame，必须添加Series.reset_index

df1 = s.loc[['F']].reset_index(name='count')
df2 = s.loc[['M']].reset_index(name='count')

或者先添加Series.reset_index：

df =  (df_n_gender_prod_cat.groupby(['Gender','prod_cat'])
                           .size()
                           .sort_index(ascending=False)
                           .reset_index(name='count'))
print (df)
  Gender     prod_cat  count
0      M  Electronics      2
1      M        Books      3
2      F  Electronics      5
3      F     Clothing      4

然后按boolean indexing过滤

df1 = df[df['Gender'].eq('F')]
df2 = df[df['Gender'].eq('M')]