Python数据分析（6）数据转换和数据聚合

心已赠人 2022-05-23 13:23 166阅读 0赞

#coding=gbk
    #数据处理
    import numpy as np
    import pandas as pd
    #轴向旋转：按照行重新调整列的顺序，或者相反
    frame=pd.DataFrame(np.arange(9).reshape(3,3),
                       index=['white','blue','black'],
                       columns=['ball','book','pen'])
    print(frame)
    #        ball  book  pen
    # white     0     1    2
    # blue      3     4    5
    # black     6     7    8
    s=frame.stack()
    print(s)
    # white  ball    0    入栈函数，将DataFrame转换成Series对象
    #        book    1
    #        pen     2
    # blue   ball    3
    #        book    4
    #        pen     5
    # black  ball    6
    #        book    7
    #        pen     8
    # dtype: int32
    f=s.unstack()   #出栈函数
    print(f)
    #        ball  book  pen
    # white     0     1    2
    # blue      3     4    5
    # black     6     7    8

在使用pandas进行数据重排时， 常要使用 stack 和 unstack 函数。stack 最好理解为堆叠， 堆积，而不要解释成入栈。

![format_png][]此图来源于：[https://www.cnblogs.com/bambipai/p/7658311.html][https_www.cnblogs.com_bambipai_p_7658311.html]

<table style="width:500px;"> 
 <tbody> 
  <tr> 
   <td>&nbsp;</td> 
   <td>store1</td> 
   <td>store2</td> 
   <td>store3</td> 
  </tr> 
  <tr> 
   <td>street1</td> 
   <td>1</td> 
   <td>2</td> 
   <td>3</td> 
  </tr> 
  <tr> 
   <td>street2</td> 
   <td>4</td> 
   <td>5</td> 
   <td>6</td> 
  </tr> 
 </tbody> 
</table>

unstack函数是从 ‘花括号结构’ 变化成 表格结构

stack函数，即进行堆叠，从表格结构转换成 ‘花括号结构’

#数据转换
    #1，删除重复数据
    frame1=pd.DataFrame({'color':['red','red','white','red','white'],
                         'value':[2,3,1,3,1]})
    print(frame1)
    #    color  value
    # 0    red      2
    # 1    red      3
    # 2  white      1
    # 3    red      3
    # 4  white      1
    print(frame1.duplicated())  #查找是否有重复元素， True为重复元素
    # 0    False
    # 1    False
    # 2    False
    # 3     True
    # 4     True
    # dtype: bool
    print(frame1[frame1.duplicated()])  #列出重复数据有哪些
    #    color  value
    # 3    red      3
    # 4  white      1
    frame2=frame1.drop_duplicates() #将重复元素删除
    print(frame2) 
    #    color  value
    # 0    red      2
    # 1    red      3
    # 2  white      1
    
    #2,离散化和面元划分
    results=[2,45,12,6,78,90,28,36,56,89]   #用于面元划分的数字
    bin=[0,25,50,75,100]    #需要划分的范围
    c=pd.cut(results,bin)   
    print(c)
    # [(0, 25], (25, 50], (0, 25], (0, 25], (75, 100], (75, 100], (25, 50], (25, 50], (50, 75], (75, 100]]
    # Categories (4, interval[int64]): [(0, 25] < (25, 50] < (50, 75] < (75, 100]]
    print(pd.value_counts(c))   #计算每个类别的元素个数
    # (75, 100]    3
    # (25, 50]     3
    # (0, 25]      3
    # (50, 75]     1
    # dtype: int64

#对DataFrame对象重新排序
    f=pd.DataFrame(np.arange(16).reshape(4,4))
    print(f)
    #     0   1   2   3
    # 0   0   1   2   3
    # 1   4   5   6   7
    # 2   8   9  10  11
    # 3  12  13  14  15
    new_order=[3,2,0,1] #设置重新排列顺序
    print(f.take(new_order))
    #     0   1   2   3
    # 3  12  13  14  15
    # 2   8   9  10  11
    # 0   0   1   2   3
    # 1   4   5   6   7

#随机抽样
    sample= np.random.randint(0,len(f),size=2)  #size随机抽取2个
    print(sample)
    # [0 1]    #相当于随机抽取了个 new_order
    print(f.take(sample))
    #    0  1  2  3
    # 0  0  1  2  3
    # 1  4  5  6  7

groupby()函数：

![20180609212918616][]

#数据聚合
    frame2=pd.DataFrame({'id':['pencil','pencil','book','ball'],
                         'color':['red','red','white','white'],
                         'price1':[45,67,90,12],
                         'price2':[89,36,12,45]})
    print(frame2)
    #   color      id  price1  price2
    # 0    red  pencil      45      89
    # 1    red  pencil      67      36
    # 2  white    book      90      12
    # 3  white    ball      12      45
    #分组函数groupby()
    group=frame2['price2'].groupby(frame2['color'])
    print(group.groups)
    # {'red': Int64Index([0, 1], dtype='int64'), 'white': Int64Index([2, 3], dtype='int64')}
    print(group.mean())
    # color
    # red      62.5    求不同颜色的价格的平均值
    # white    28.5
    # Name: price2, dtype: float64

#2等级分组
    print(frame2['price2'].groupby([frame2['color'],frame2['id']]).sum())
    # color  id                    看清楚[frame2['color'],frame2['id']
    # red    pencil    125        将color 和 id 作为分组的依据
    # white  ball       45
    #        book       12
    # Name: price2, dtype: int64
    print(frame2[['price1','price2']].groupby(frame2['id']).sum())
    #         price1  price2
    # id                        看清楚frame2[['price1','price2']]，中括号顺序
    # ball        12      45
    # book        90      12
    # pencil     112     125

--------------------

![1240][]

[format_png]: /images/20220523/352d8525a826412e9bdf771d6f142461.png
[https_www.cnblogs.com_bambipai_p_7658311.html]: https://www.cnblogs.com/bambipai/p/7658311.html
[20180609212918616]: /images/20220523/afb108df73144d719b9ce5e0c8662f98.png
[1240]: /images/20220523/196ee868e7e84c0695fdd86df6d46bdf.png