熊猫 分组

import pandas as pd
import numpy as np

group by : split apply combine

dic_gpa = {‘class’:[‘A’,‘B’,‘A’,‘B’],
‘sname’:[‘tom’,‘kite’,‘tom’,‘hanmeimei’],
‘math’:(np.random.rand(4)*100).round(),
‘english’:(np.random.rand(4)*100).round()
}

df = pd.DataFrame(dic_gpa)

smean_byclass = df.groupby([‘class’]).mean()
smean_byname = df.groupby([‘sname’]).mean()
lambda_name = df.groupby(lambda x:‘even class’ if x%2==0 else ‘odd class’).mean()

agg_byClassSname = df.groupby([‘class’,‘sname’]).agg([np.mean,np.min])
agg_byClassSname_different = df.groupby([‘class’,‘sname’]).agg({‘math’:np.mean,‘english’:np.max})

mean_minus = df.groupby(‘class’).apply(lambda x:x[‘math’].mean()-x[‘english’].mean())

transformer = df.groupby(‘class’).transform(np.mean)

combine_concat = pd.concat([df,transformer],axis=1)

print(df)
print(’*’*100)
#print(smean_byclass)
print(smean_byclass)

combine_merge = pd.merge(df,smean_byclass,how=‘left’,on=[‘class’])

print(combine_merge)