Python Pandas的分组运算

分组运算分为三步：拆分 分组 运算
拆分：根据什么进行分组
应用：每个分组进行什么样的运算
合并：把每个分组的计算结果合并起来

df = pd.DataFrame({'key1': ['a', 'a', 'b', 'b', 'a'],
                   'key2': ['one', 'two', 'one', 'two', 'one'],
                   'data1': np.random.randint(1, 10, 5),
                   'data2': np.random.randint(1, 10, 5)})
print(df)

  key1 key2  data1  data2
0    a  one      1      7
1    a  two      5      4
2    b  one      8      9
3    b  two      5      7
4    a  one      2      4

将data1与key1进行分组，进行相加的操作

print(df['data1'].groupby(df['key1']).sum())

key1
a     8
b    13
Name: data1, dtype: int32

对指定的key进行分组，相加

key = [1, 2, 1, 1, 2]
print(df['data1'].groupby(key).sum())

1    20
2     6
Name: data1, dtype: int32

多层分组

print(df['data1'].groupby([df['key1'], df['key2']]).sum())

key1  key2
a     one     4
      two     7
b     one     5
      two     1

mean = df.groupby(['key1', 'key2']).sum()['data1']
print(mean.unstack())

key2  one  two
key1          
a      15    9
b       6    4

for name, group in df.groupby('key1'):
    print(name)
    print(group)

a
  key1 key2  data1  data2
0    a  one      5      1
1    a  two      7      6
4    a  one      8      1
b
  key1 key2  data1  data2
2    b  one      6      4
3    b  two      8      7

按列进行分组求和

mapping = {'a': 'red', 'b': 'red', 'c': 'blue', 'd': 'orange', 'e': 'blue'}
grouped = df.groupby(mapping, axis=1)
print(grouped.sum())

       blue  orange  red
Alice    14       3   14
Bob      15       2   11
Candy    13       7    8
Dark     13       5   13
Emily     7       2   12

查看分组的情况

blue      2
orange    1
red       2

多级索引
行索引一级，列索引两级

columns = pd.MultiIndex.from_arrays([['China', 'USA', 'China', 'US', 'China'],
                                     ['A', 'A', 'B', 'C', 'B']], names=['country', 'index'])
df = pd.DataFrame(np.random.randint(1, 10, (5, 5)), columns=columns)
print(df)

country China USA China US China
index       A   A     B  C     B
0           3   8     4  4     3
1           6   1     6  6     2
2           9   8     7  5     5
3           2   1     1  4     3
4           2   3     5  5     6

对country的列索引进行分组

country  China  US  USA
0           16   2    7
1           13   5    8
2           11   8    1
3           23   9    6
4           13   7    3