Pandas GroupBy 使用教程

实例 1 将分组后的字符拼接

import pandas as pd
df=pd.DataFrame({
    'user_id':[1,2,1,3,3],
    'content_id':[1,1,2,2,2],
    'tag':['cool','nice','clever','clever','not-bad']
})
df

将df按content_id分组，然后将每组的tag用逗号拼接

df.groupby('content_id')['tag'].apply(lambda x:','.join(x)).to_frame()

实例2 统计每个content_id有多少个不同的用户

import pandas as pd

df = pd.DataFrame({
    'user_id':[1,2,1,3,3,],
    'content_id':[1,1,2,2,2],
    'tag':['cool','nice','clever','clever','not-bad']
})

df.groupby("content_id")["user_id"].nunique().to_frame()

实例3 分组结果排序

import pandas as pd

df = pd.DataFrame({
    'value':[20.45,22.89,32.12,111.22,33.22,100.00,99.99],
    'product':['table','chair','chair','mobile phone','table','mobile phone','table']
})
df

df1 = df.groupby('product')['value'].sum().to_frame().reset_index()
df1

按产品product分组后，然后value求和：

df2 = df.groupby('product')['value'].sum().to_frame().reset_index().sort_values(by='value')
df2

实例4 分组大小绘图

import pandas as pd
import matplotlib.pyplot as plt

df = pd.DataFrame({
    'value':[20.45,22.89,32.12,111.22,33.22,100.00,99.99],
    'product':['table','chair','chair','mobile phone','table','mobile phone','table']
})
df

plt.clf()
df.groupby('product').size().plot(kind='bar')
plt.show()

实例5 分组求和绘图

import pandas as pd
import matplotlib.pyplot as plt

df = pd.DataFrame({
    'value':[20.45,22.89,32.12,111.22,33.22,100.00,99.99],
    'product':['table','chair','chair','mobile phone','table','mobile phone','table']
})
df

plt.clf()
df.groupby('product').sum().plot(kind='bar')
plt.show()

实例 6 使用agg函数

import pandas as pd

df = pd.DataFrame({
    'value':[20.45,22.89,32.12,111.22,33.22,100.00,99.99],
    'product':['table','chair','chair','mobile phone','table','mobile phone','table']
})

grouped_df = df.groupby('product').agg({'value':['min','max','mean']})
grouped_df

grouped_df.columns = ['_'.join(col).strip() for col in grouped_df.columns.values]
grouped_df = grouped_df.reset_index()
grouped_df

实例7 遍历分组

for key,group_df in df.groupby('product'):
    print("the group for product '{}' has {} rows".format(key,len(group_df)))  

the group for product 'chair' has 2 rows
the group for product 'mobile phone' has 2 rows
the group for product 'table' has 3 rows

源代码：Python008-Pandas GroupBy 使用教程.ipynb