pandas之聚合运算

通过聚合运算可以得到我们比较感兴趣的数据以方便处理

 import pandas as pd
 import numpy as np

 # 先创建一组数据表DataFrame
 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)})
 # 这组数据的key1列与key2列有相同的有各自相同的索引行

 df.groupby('key1').describe()
 # 通过这一行我们可以看到关于key1列分组后的一些计算结果，
 # count  mean std min 25% 50% 75% max

 grouped = df.groupby('key1')
 # 它返回的是DataFrom对象

 def peak_range(s):
     print(type(s))
     return s.max() - s.min()

 # 接下来就是看看聚合的应用了
 grouped.agg(['std','mean','sum',('range',peak_range)])
 # 这个函数聚合了df.groupby('key1').describe()所展现的一些计算结果
 # ('range',peak_range)中的range是peak_range函数计算结果的列名

 # 通过创建字典进行聚合自己想要的部分数据
 d = {'data1':'mean',
     'data2':'sum'}
 grouped.agg(d)

 d = {'data1':['mean',('range',peak_range)],
     'data2':'sum'}
 grouped.agg(d)
 grouped.agg(d).reset_index()  # 不把key1作为索引来处理
 df.groupby('key1',as_index=False).agg(d)  # 与上一行效果一样

 import pandas as pd
 import numpy as np

 # 先创建一个DataFrame
 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)})
 # 合并方式一
 k1_mean = df.groupby('key1').mean().add_prefix('mean_')
 # 将key1分组取均值，并将data1与data2之前加上mean_
 pd.merge(df, k1_mean, left_on='key1',right_index=True)  # 通过merge合并

 k1_mean = df.groupby('key1').transform(np.mean).add_prefix('mean_')
 # 它通过分组求平均，保持原来的索引位置与行数
 # 这样可以通过下面的代码实现合并
 df[k1_mean.columns] = k1_mean

 import pandas as pd
 import numpy as np

 df = pd.DataFrame(np.random.randint(1,10,(5,5)),
                  columns=['a','b','c','d','e'],
                  index=['Alice','Bob','Candy','Dark','Emily'])

 def demean(s):
     return s - s.mean()

 key = ['one','one','two','one','two']
 demeaned = df.groupby(key).transform(demean)

 demeaned.groupby(key).mean()  # 输出的值都是0或者接近0

 import pandas as pd
 import numpy as np

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

 def top(g, n=2, column='data1'):
     return g.sort_values(by=column,ascending=False)[:n]

 df.groupby('key1').apply(top, n=3, column='data2')
 # 分组后只对data2列的值排序取出前三行

 import pandas as pd
 import numpy as np

 # 下面的例子是填充NaN值的方法

 states = ['Ohio','New York','Vermont','Florida',
          'Oregon','Nevada','California','Idaho']
 group_key = ['East'] * 4 + ['West'] * 4
 data = pd.Series(np.random.randn(8), index=states)
 data[['Vermont','Nevada','Idaho']] = np.nan
 # 输出
 Ohio          0.133410
 New York      2.147483
 Vermont            NaN
 Florida      -0.608754
 Oregon        0.978375
 Nevada             NaN
 California   -1.297183
 Idaho              NaN
 dtype: float64

 data.groupby(group_key).mean()
 # 输出
 East    0.557380
 West   -0.159404
 dtype: float64

 data.groupby(group_key).apply(lambda g: g.fillna(g.mean()))
 # 输出
 Ohio          0.133410
 New York      2.147483
 Vermont       0.557380
 Florida      -0.608754
 Oregon        0.978375
 Nevada       -0.159404
 California   -1.297183
 Idaho        -0.159404
 dtype: float64