Pandas：GroupBy--对分组进行迭代

import pandas as pd
import numpy as np
from pandas import Series,DataFrame
df = DataFrame({'key1':['a','a','b','b','a'],
               'key2':['one','two','one','two','one'],
               'data1':np.random.randn(5),
               'data2':np.random.randn(5)})
print(df)

      data1     data2 key1 key2
0  0.001573  1.348387    a  one
1 -0.423522  0.884686    a  two
2 -0.942151 -0.672910    b  one
3 -0.720580  0.724431    b  two
4  0.262283 -0.251035    a  one

1.GroupBy对象支持迭代，可以产生一组二元元组(由分组名和数据块组成)

for name,group in df.groupby('key1'):
    print(name)
    print(group,end='\n-------------------------------\n')

a
      data1     data2 key1 key2
0  0.001573  1.348387    a  one
1 -0.423522  0.884686    a  two
4  0.262283 -0.251035    a  one
-------------------------------
b
      data1     data2 key1 key2
2 -0.942151 -0.672910    b  one
3 -0.720580  0.724431    b  two
-------------------------------

2.多重键分组进行迭代

for (key1,key2),group in df.groupby(['key1','key2']):
    print(key1,key2)
    print(group,end='\n-------------------------------\n')

a one
      data1     data2 key1 key2
0  0.001573  1.348387    a  one
4  0.262283 -0.251035    a  one
-------------------------------
a two
      data1     data2 key1 key2
1 -0.423522  0.884686    a  two
-------------------------------
b one
      data1    data2 key1 key2
2 -0.942151 -0.67291    b  one
-------------------------------
b two
     data1     data2 key1 key2
3 -0.72058  0.724431    b  two
-------------------------------

3.将GroupBy对象转换为字典

分组键就是字典的键，分组就是字典的值

pieces = dict(list(df.groupby('key1')))
print(pieces['b'])

      data1     data2 key1 key2
2 -0.942151 -0.672910    b  one
3 -0.720580  0.724431    b  two

4.按列的类型(dtype)进行分组

grouped = df.groupby(df.dtypes,axis=1)
dict(list(grouped))

{dtype('float64'):       data1     data2
 0  0.001573  1.348387
 1 -0.423522  0.884686
 2 -0.942151 -0.672910
 3 -0.720580  0.724431
 4  0.262283 -0.251035, dtype('O'):   key1 key2
 0    a  one
 1    a  two
 2    b  one
 3    b  two
 4    a  one}