【Data Mining】机器学习三剑客之Pandas常用算法总结下

Data mining 同时被 2 个专栏收录
7 篇文章 0 订阅

前言

请先看:
机器学习三剑客之Numpy常用算法总结
机器学习三剑客之Pandas常用算法总结上

设置值

import pandas as pd
import numpy as np

dates = pd.date_range('20191222', periods=3)
df = pd.DataFrame(np.arange(12).reshape((3, 4)), index=dates, columns=['A', 'B', 'C', 'D'])
print(df)
"""
            A  B   C   D
2019-12-22  0  1   2   3
2019-12-23  4  5   6   7
2019-12-24  8  9  10  11
"""
# index
df.iloc[1, 1] = 1111
print(df)
"""
            A     B   C   D
2019-12-22  0     1   2   3
2019-12-23  4  1111   6   7
2019-12-24  8     9  10  11
"""
# label
df.loc['20191224', 'C'] = 2222
print(df)
"""
            A     B     C   D
2019-12-22  0     1     2   3
2019-12-23  4  1111     6   7
2019-12-24  8     9  2222  11
"""
# mix
df.ix['20191222', 1] = 3333
print(df)
"""
            A     B     C   D
2019-12-22  0  3333     2   3
2019-12-23  4  1111     6   7
2019-12-24  8     9  2222  11
"""




dates2 = pd.date_range('20191222', periods=6)
df2 = pd.DataFrame(np.arange(24).reshape((6, 4)), index=dates2, columns=['A', 'B', 'C', 'D'])
print(df2)
"""
             A   B   C   D
2019-12-22   0   1   2   3
2019-12-23   4   5   6   7
2019-12-24   8   9  10  11
2019-12-25  12  13  14  15
2019-12-26  16  17  18  19
2019-12-27  20  21  22  23
"""
df2[df2.A > 12] = 0
print(df2)
df2.A[df2.A > 8] = 0
print(df2)
df2.B[df2.A > 2] = 0
print(df2)
# batch processing
"""
             A   B   C   D
2019-12-22   0   1   2   3
2019-12-23   4   5   6   7
2019-12-24   8   9  10  11
2019-12-25  12  13  14  15
2019-12-26   0   0   0   0
2019-12-27   0   0   0   0
            A   B   C   D
2019-12-22  0   1   2   3
2019-12-23  4   5   6   7
2019-12-24  8   9  10  11
2019-12-25  0  13  14  15
2019-12-26  0   0   0   0
2019-12-27  0   0   0   0
            A   B   C   D
2019-12-22  0   1   2   3
2019-12-23  4   0   6   7
2019-12-24  8   0  10  11
2019-12-25  0  13  14  15
2019-12-26  0   0   0   0
2019-12-27  0   0   0   0
"""
# add a colomn
df2['F'] = np.nan
print df2
"""
            A   B   C   D   F
2019-12-22  0   1   2   3 NaN
2019-12-23  4   0   6   7 NaN
2019-12-24  8   0  10  11 NaN
2019-12-25  0  13  14  15 NaN
2019-12-26  0   0   0   0 NaN
2019-12-27  0   0   0   0 NaN
"""
# add a colomn using Series
df2['E'] = pd.Series([1, 2, 3, 4, 5, 6], index=dates2)
print df2
"""
            A   B   C   D   F  E
2019-12-22  0   1   2   3 NaN  1
2019-12-23  4   0   6   7 NaN  2
2019-12-24  8   0  10  11 NaN  3
2019-12-25  0  13  14  15 NaN  4
2019-12-26  0   0   0   0 NaN  5
2019-12-27  0   0   0   0 NaN  6
"""

一些总结说明

  • 首先第一种方法,毋庸置疑,猜都能猜到就是之前的选取值的方式之后对于选取的值进行赋值即可,主要有index,label和mix
  • 对于数字数据,通过比较大小来批量处理数据,可以好几行好几列一起处理,也可以单独列或者单独行处理
  • 设置值还包括添加新的一个colomn,也就是添加一个新的series,如果添加的为标量则可以直接一个数字一起赋值,但如果输入过于复杂,我个人建议都采用Series的方式进行添加,所以这就最后归因于你是否掌握了series,方式为dataframe[‘colomn name’] = pd.Series([])

处理缺省值

# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np

dates = pd.date_range('20191222', periods=4)
df = pd.DataFrame(np.arange(16).reshape((4, 4)), index=dates, columns=['A', 'B', 'C', 'D'])
print(df)
"""
             A   B   C   D
2019-12-22   0   1   2   3
2019-12-23   4   5   6   7
2019-12-24   8   9  10  11
2019-12-25  12  13  14  15
"""
# add nan
df.iloc[0, 1] = np.nan
df.iloc[2, 2] = np.nan
print(df)
# drop na (index or colomn)
print(df.dropna(axis=0, how='any'))
print(df.dropna(axis=1, how='any'))
"""
             A     B     C   D
2019-12-22   0   NaN   2.0   3
2019-12-23   4   5.0   6.0   7
2019-12-24   8   9.0   NaN  11
2019-12-25  12  13.0  14.0  15
             A     B     C   D
2019-12-23   4   5.0   6.0   7
2019-12-25  12  13.0  14.0  15
             A   D
2019-12-22   0   3
2019-12-23   4   7
2019-12-24   8  11
2019-12-25  12  15
"""

# 前后对比
print(df.dropna(axis=1, how='all'))
"""
             A     B     C   D
2019-12-22   0   NaN   2.0   3
2019-12-23   4   5.0   6.0   7
2019-12-24   8   9.0   NaN  11
2019-12-25  12  13.0  14.0  15
"""
df.iloc[1, 1] = np.nan
df.iloc[2, 1] = np.nan
df.iloc[3, 1] = np.nan
print(df)
print(df.dropna(axis=1, how='all'))
"""
             A   B     C   D
2019-12-22   0 NaN   2.0   3
2019-12-23   4 NaN   6.0   7
2019-12-24   8 NaN   NaN  11
2019-12-25  12 NaN  14.0  15
             A     C   D
2019-12-22   0   2.0   3
2019-12-23   4   6.0   7
2019-12-24   8   NaN  11
2019-12-25  12  14.0  15
"""
# fillna has many parameters
# please see details
# 把nan复制成value
print(df.fillna(value=222.22))
"""
             A       B       C   D
2019-12-22   0  222.22    2.00   3
2019-12-23   4  222.22    6.00   7
2019-12-24   8  222.22  222.22  11
2019-12-25  12  222.22   14.00  15
"""
# judge nan
# 一般表格很大的时候用
# nan return True
print(df.isnull())
# 如果有一个值为df.isnull 中有一个为True,则返回True
# 一起使用,则为判断这个大的dataframe中是否含有nan
print(np.any(df.isnull()))

print(np.any(df.isnull()) == True)

"""
                A     B      C      D
2019-12-22  False  True  False  False
2019-12-23  False  True  False  False
2019-12-24  False  True   True  False
2019-12-25  False  True  False  False
True
True
"""

  • 删除nan,all全部为nan则删,any有一个nan则删,删index or colomn看axis
  • 填充nan一个value
  • 超级大的dataframe判断是否有nan

合并concat

# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np

df1 = pd.DataFrame(np.ones((2, 3))*0, columns=['a', 'b', 'c'])
df2 = pd.DataFrame(np.ones((2, 3))*1, columns=['a', 'b', 'c'])
df3 = pd.DataFrame(np.ones((2, 3))*22, columns=['a', 'b', 'c'])
print(df1)
print(df2)
print(df3)
"""
     a    b    c
0  0.0  0.0  0.0
1  0.0  0.0  0.0
     a    b    c
0  1.0  1.0  1.0
1  1.0  1.0  1.0
      a     b     c
0  22.0  22.0  22.0
1  22.0  22.0  22.0
"""
# 三个进行合并 index也是最初的组合而已,此处axis=0为vertical合并
df_vertical = pd.concat([df1, df2, df3], axis=0)
print(df_vertical)
"""
      a     b     c
0   0.0   0.0   0.0
1   0.0   0.0   0.0
0   1.0   1.0   1.0
1   1.0   1.0   1.0
0  22.0  22.0  22.0
1  22.0  22.0  22.0
"""
#进行index的重新洗牌
df_vertical_rightindex = pd.concat([df1, df2, df3], axis=0, ignore_index=True)
print(df_vertical_rightindex)
"""
      a     b     c
0   0.0   0.0   0.0
1   0.0   0.0   0.0
2   1.0   1.0   1.0
3   1.0   1.0   1.0
4  22.0  22.0  22.0
5  22.0  22.0  22.0
"""
# horizontal
df_horizontal = pd.concat([df1, df2, df3], axis=1)
print(df_horizontal)
"""
     a    b    c    a    b    c     a     b     c
0  0.0  0.0  0.0  1.0  1.0  1.0  22.0  22.0  22.0
1  0.0  0.0  0.0  1.0  1.0  1.0  22.0  22.0  22.0
"""
df_horizontal_rightindex = pd.concat([df1, df2, df3], axis=1, ignore_index=True)
print(df_horizontal_rightindex)
"""
     0    1    2    3    4    5     6     7     8
0  0.0  0.0  0.0  1.0  1.0  1.0  22.0  22.0  22.0
1  0.0  0.0  0.0  1.0  1.0  1.0  22.0  22.0  22.0
"""
df_horizontal_rightindex.rename(
    columns={0: 'a', 1: 'b', 2: 'c', 3: 'd', 4: 'e',
             5: 'f', 6: 'g', 7: 'h', 8: 'i', 9: 'g'}, inplace=True)
print(df_horizontal_rightindex)
"""
     a    b    c    d    e    f     g     h     i
0  0.0  0.0  0.0  1.0  1.0  1.0  22.0  22.0  22.0
1  0.0  0.0  0.0  1.0  1.0  1.0  22.0  22.0  22.0
"""

一些总结:

  • axis=0 为纵向合并,axis=1为横向合并
  • 纵向合并时,想重新排序index,采用ignore_index=True参数即可
  • 横向合并时,相当于把column作为index,所以采用ignore_index=True操作最终结果是所有的column变为0,1,2,3…,所以需要rename(columns={})重新命名即可

  # different colomns label and index label,need to use 'join' and 'join_axes' parameters
df4 = pd.DataFrame(np.ones((3, 4))*0, columns=['a', 'b', 'c', 'd'], index=[1, 2, 3])
df5 = pd.DataFrame(np.ones((3, 4))*1, columns=['b', 'c', 'd', 'e'], index=[2, 3, 4])
print(df4)
print(df5)
"""
     a    b    c    d
1  0.0  0.0  0.0  0.0
2  0.0  0.0  0.0  0.0
3  0.0  0.0  0.0  0.0

     b    c    d    e
2  1.0  1.0  1.0  1.0
3  1.0  1.0  1.0  1.0
4  1.0  1.0  1.0  1.0
"""
# 不存在的用nan填充
df_outer = pd.concat([df4, df5], join='outer', ignore_index=True)
print(df_outer)
"""
     a    b    c    d    e
0  0.0  0.0  0.0  0.0  NaN
1  0.0  0.0  0.0  0.0  NaN
2  0.0  0.0  0.0  0.0  NaN
3  NaN  1.0  1.0  1.0  1.0
4  NaN  1.0  1.0  1.0  1.0
5  NaN  1.0  1.0  1.0  1.0
"""
df_inner = pd.concat([df4, df5], join='inner')
print(df_inner)
"""
     b    c    d
1  0.0  0.0  0.0
2  0.0  0.0  0.0
3  0.0  0.0  0.0
2  1.0  1.0  1.0
3  1.0  1.0  1.0
4  1.0  1.0  1.0
"""
df_inner_index = pd.concat([df4, df5], join='inner', ignore_index=True)
print(df_inner_index)
"""
     b    c    d
0  0.0  0.0  0.0
1  0.0  0.0  0.0
2  0.0  0.0  0.0
3  1.0  1.0  1.0
4  1.0  1.0  1.0
5  1.0  1.0  1.0
"""


"""
原df4,df5 便于观看 
     a    b    c    d
1  0.0  0.0  0.0  0.0
2  0.0  0.0  0.0  0.0
3  0.0  0.0  0.0  0.0

     b    c    d    e
2  1.0  1.0  1.0  1.0
3  1.0  1.0  1.0  1.0
4  1.0  1.0  1.0  1.0
"""
# 不存在的用nan填充
df_h_different_index = pd.concat([df4, df5], axis=1)
print(df_h_different_index)
"""
     a    b    c    d    b    c    d    e
1  0.0  0.0  0.0  0.0  NaN  NaN  NaN  NaN
2  0.0  0.0  0.0  0.0  1.0  1.0  1.0  1.0
3  0.0  0.0  0.0  0.0  1.0  1.0  1.0  1.0
4  NaN  NaN  NaN  NaN  1.0  1.0  1.0  1.0
"""
#用df4的index作为合并之后的index,所以不是df4的index部分删掉(4删掉)
df_h_different_index_df4 = pd.concat([df4, df5], axis=1, join_axes=[df4.index])
print(df_h_different_index_df4)
"""
     a    b    c    d    b    c    d    e
1  0.0  0.0  0.0  0.0  NaN  NaN  NaN  NaN
2  0.0  0.0  0.0  0.0  1.0  1.0  1.0  1.0
3  0.0  0.0  0.0  0.0  1.0  1.0  1.0  1.0
"""

df_h_different_index_df5 = pd.concat([df4, df5], axis=1, join_axes=[df5.index])
print(df_h_different_index_df5)
"""
     a    b    c    d    b    c    d    e
2  0.0  0.0  0.0  0.0  1.0  1.0  1.0  1.0
3  0.0  0.0  0.0  0.0  1.0  1.0  1.0  1.0
4  NaN  NaN  NaN  NaN  1.0  1.0  1.0  1.0
""" 

一些总结:

  • 当index和column label不相同时合并
  • 纵向合并用join=[‘outer’,‘inner’],横向合并用join_axes=[df.index]
  • 当纵向时,outer为全留下,没有值的地方为nan,inner为只要相同的colomn label处
  • 当横向时,不采用join_axes和join='outer’效果相似,如果采用join_axes则以对应的index为最后的index,其余删掉

合并append

import numpy as np 
import pandas as pd 
df1 = pd.DataFrame(np.ones((2, 3))*0, columns=['a', 'b', 'c']) 
df2 = pd.DataFrame(np.ones((2, 3))*11, columns=['a', 'b', 'c'])  
df3 = pd.DataFrame(np.ones((2, 3))*2, columns=['a', 'b', 'c']) 
print(df1) 
print(df2) 
print(df3) 
""" 
     a    b    c 
0  0.0  0.0  0.0 
1  0.0  0.0  0.0 
      a     b     c 
0  11.0  11.0  11.0 
1  11.0  11.0  11.0  
     a    b    c 
0  2.0  2.0  2.0 
1  2.0  2.0  2.0  
""" 
df_append1 = df1.append(df2, ignore_index=True)       
print(df_append1)                     
"""             
      a     b     c                                                          
0   0.0   0.0   0.0                                                     
1   0.0   0.0   0.0                                                           
2  11.0  11.0 11.0    
3  11.0  11.0  11.0 
"""          
df_append2 = df1.append([df2, df3], ignore_index=True)      
print(df_append2)                
"""      
      a     b     c    
0   0.0   0.0   0.0                                                     
1   0.0   0.0   0.0                                                       
2  11.0  11.0  11.0                                                         
3  11.0  11.0  11.0   
4   2.0   2.0   2.0      
5   2.0   2.0   2.0                          
"""                                           
s1 = pd.Series([12, 24, 33], index=['a', 'b', 'c'])  
df_appned_s = df1.append(s1, ignore_index=True)  
print(df_appned_s) 
"""   
      a     b     c 
0   0.0   0.0   0.0 
1   0.0   0.0   0.0 
2  12.0  24.0  33.0
""" 

一些总结:

  • append是python中list常用的功能,这里append相当于dataframe在纵向方向上加数据
  • series相当于dataframe的一行,append一个series相当于在下面加上了一行

合并merge

1. on对应的column中的value完全相同,并且只针对一个column

# -*- coding: utf-8 -*-
import pandas as pd

left = pd.DataFrame({'connect': ['con0', 'con1', 'con2', 'con3'],
                             'A': ['A0', 'A1', 'A2', 'A3'],
                             'B': ['B0', 'B1', 'B2', 'B3']})
right = pd.DataFrame({'connect' : ['con0', 'con1', 'con2', 'con3'],
                            'C': ['C0', 'C1', 'C2', 'C3'],
                            'D': ['D0', 'D1', 'D2', 'D3']})

print(left)
print(right)
"""
    A   B connect
0  A0  B0    con0
1  A1  B1    con1
2  A2  B2    con2
3  A3  B3    con3
    C   D connect
0  C0  D0    con0
1  C1  D1    con1
2  C2  D2    con2
3  C3  D3    con3
"""


# 两个dataframe通过相同的column,merge在一起
# column label为connect上的元素全部一样的一般情况
merge_l_r_same = pd.merge(left=left, right=right, on='connect')
print(merge_l_r_same)
merge_l_r_same = pd.merge(left=right, right=left, on='connect')
print(merge_l_r_same)
"""
    A   B connect   C   D
0  A0  B0    con0  C0  D0
1  A1  B1    con1  C1  D1
2  A2  B2    con2  C2  D2
3  A3  B3    con3  C3  D3

    C   D connect   A   B
0  C0  D0    con0  A0  B0
1  C1  D1    con1  A1  B1
2  C2  D2    con2  A2  B2
3  C3  D3    con3  A3  B3

通过这个对比可以明白参数on, 以及left, right的功能
left,right即通过on参数指定的column来进行合并而
left对应的dataframe剩下的column则全都依次放在left
right对应的dataframe剩下的column则全都依次放在right
"""

2. on针对多个column,并且对应的column中的value不相同

# on多个column 并且每个column label对应的value有重叠但不完全相同.

left = pd.DataFrame({'key1': ['K1', 'K0', 'K1', 'K2'],
                      'key2': ['K0', 'K0', 'K1', 'K1'],
                      'A': ['A0', 'A1', 'A2', 'A3'],
                      'B': ['B0', 'B1', 'B2', 'B3']})
right = pd.DataFrame({'key1': ['K0', 'K0', 'K1', 'K1'],
                       'key2': ['K0', 'K1', 'K1', 'K1'],
                       'C': ['C0', 'C1', 'C2', 'C3'],
                       'D': ['D0', 'D1', 'D2', 'D3']})
print(left)
print(right)
"""
    A   B key1 key2
0  A0  B0   K1   K0
1  A1  B1   K0   K0
2  A2  B2   K1   K1
3  A3  B3   K2   K1

    C   D key1 key2
0  C0  D0   K0   K0
1  C1  D1   K0   K1
2  C2  D2   K1   K1
3  C3  D3   K1   K1
"""
# on对应于多个column时需要采用List
# how=['inner', 'outer', 'left', 'right']
# default = 'inner'
merge_l_r_different = pd.merge(left, right, on=['key1', 'key2'])
print(merge_l_r_different)
"""
    A   B key1 key2   C   D
0  A1  B1   K0   K0  C0  D0
1  A2  B2   K1   K1  C2  D2
2  A2  B2   K1   K1  C3  D3
"""
merge_inner = pd.merge(left, right, on=['key1', 'key2'], how='inner')
print(merge_inner)
"""
    A   B key1 key2   C   D
0  A1  B1   K0   K0  C0  D0
1  A2  B2   K1   K1  C2  D2
2  A2  B2   K1   K1  C3  D3
"""

"""
便于观看
left:
    A   B key1 key2
0  A0  B0   K1   K0
1  A1  B1   K0   K0
2  A2  B2   K1   K1
3  A3  B3   K2   K1
right:
    C   D key1 key2
0  C0  D0   K0   K0
1  C1  D1   K0   K1
2  C2  D2   K1   K1
3  C3  D3   K1   K1
"""

merge_outer = pd.merge(left, right, on=['key1', 'key2'], how='outer')
print(merge_outer)
"""
     A    B key1 key2    C    D
0   A0   B0   K1   K0  NaN  NaN
1   A1   B1   K0   K0   C0   D0
2   A2   B2   K1   K1   C2   D2
3   A2   B2   K1   K1   C3   D3
4   A3   B3   K2   K1  NaN  NaN
5  NaN  NaN   K0   K1   C1   D1
"""
merge_left = pd.merge(left, right, on=['key1', 'key2'], how='left')
print(merge_left)
"""
    A   B key1 key2    C    D
0  A0  B0   K1   K0  NaN  NaN
1  A1  B1   K0   K0   C0   D0
2  A2  B2   K1   K1   C2   D2
3  A2  B2   K1   K1   C3   D3
4  A3  B3   K2   K1  NaN  NaN
"""

"""
便于观看
left:
    A   B key1 key2
0  A0  B0   K1   K0
1  A1  B1   K0   K0
2  A2  B2   K1   K1
3  A3  B3   K2   K1
right:
    C   D key1 key2
0  C0  D0   K0   K0
1  C1  D1   K0   K1
2  C2  D2   K1   K1
3  C3  D3   K1   K1
"""
merge_right = pd.merge(left, right, on=['key1', 'key2'], how='right')
print(merge_right)
"""
     A    B key1 key2   C   D
0   A1   B1   K0   K0  C0  D0
1   A2   B2   K1   K1  C2  D2
2   A2   B2   K1   K1  C3  D3
3  NaN  NaN   K0   K1  C1  D1
"""

这里主要是考察merge函数中的how参数的用法,以及对应的operations
这里举两个例子主要分析

"""
    A   B key1 key2
0  A0  B0   K1   K0
1  A1  B1   K0   K0
2  A2  B2   K1   K1
3  A3  B3   K2   K1

    C   D key1 key2
0  C0  D0   K0   K0
1  C1  D1   K0   K1
2  C2  D2   K1   K1
3  C3  D3   K1   K1
"""
result:
"""
    A   B key1 key2   C   D
0  A1  B1   K0   K0  C0  D0
1  A2  B2   K1   K1  C2  D2
2  A2  B2   K1   K1  C3  D3
"""
进行on=['key1, key2'], how='inner'
请试着和我一起分析:
首先因为通过key1和key2两个column进行合并,观察两列中的[key1,key2]组合完全相同的行
可以明显观察到的是 [K0,K0]以及[K1,K1]相同,因为采用inner方式所以只考虑相同的行即可,
首先取出[K0,K0]中的left的元素放置于左边,对应的right的value放置于右边,接下来一样的操作
对应left中的[K1, K1]有两个right对应,首先先将[K1, K1]的left的元素放置于左边,之后去取
right的中的元素放置于右面,而对于right中的[K1, K1]有两个放置于右边,所以左边还需补一组
一模一样的left中对应的数据放置于左面,即可。
如果how='outer'的话则全部的columns都留下,只是left和right中没有对应的部分,为nan即可,可返回到代码部分重复观看,即可明白。
"""
    A   B key1 key2    C    D
0  A0  B0   K1   K0  NaN  NaN
1  A1  B1   K0   K0   C0   D0
2  A2  B2   K1   K1   C2   D2
3  A2  B2   K1   K1   C3   D3
4  A3  B3   K2   K1  NaN  NaN
"""
how='left'的意思才估计也能猜到了,就是主要针对于left的所有的[key1, key2]组合,right只是
迎合left的结构。分析如下:
首先left中的[K1,K0]对应于的数据放在左面,而right中不存在则为nan nan,之后[K0,K0]存在对应的right元素,[K1,K1]和上面的分析基本一致,最后到[K2, K1][K1, K0]一个意思。
大家可以自行根据以上分析来分析,'outer''right'情况

3.indicator

import pandas as pd

# the parameter of indicator
left = pd.DataFrame({'key': [0, 1], 'left': ['a', 'b']})
right = pd.DataFrame({'key': [1, 2, 2], 'right': [2, 2, 2]})
print(left)
print(right)
"""
   key left
0    0    a
1    1    b
   key  right
0    1      2
1    2      2
2    2      2
"""
res_indicator = pd.merge(left=left, right=right, on='key', how='outer', indicator=True)
print(res_indicator)
"""
   key left  right      _merge
0    0    a    NaN   left_only
1    1    b    2.0        both
2    2  NaN    2.0  right_only
3    2  NaN    2.0  right_only
"""
# give indicater a name that is "dict_name"
# the default name is "_merge"
res_indicator2 = pd.merge(left, right, on='key', how='outer', indicator="idct_name")
print(res_indicator2)
"""
   key left  right   idct_name
0    0    a    NaN   left_only
1    1    b    2.0        both
2    2  NaN    2.0  right_only
3    2  NaN    2.0  right_only
"""

merge的parameter “indicator”主要作用为显示出最终的left和right一起merge的结果,为left_only:这一index只有left有value,而right无value,both则都有,right_only和left_only相似。

利用index进行merge(left_index,right_index),配合使用(left_on, right_on)


# define the indexes of left and right again
left = pd.DataFrame({'A': ['A0', 'A1', 'A2'],
                     'B': ['B0', 'B1', 'B2']},
                     index=['K0', 'K1', 'K2'])

right = pd.DataFrame({'C': ['C0', 'C2', 'C3'],
                      'D': ['D0', 'D2', 'D3']},
                     index=['K0', 'K2', 'K3'])
print(left)
print(right)
"""
     A   B
K0  A0  B0
K1  A1  B1
K2  A2  B2
     C   D
K0  C0  D0
K2  C2  D2
K3  C3  D3
"""
# classification of (left_index, right_index)
res = pd.merge(left, right, left_index=True, right_index=True, how='outer')
print res
"""
      A    B    C    D
K0   A0   B0   C0   D0
K1   A1   B1  NaN  NaN
K2   A2   B2   C2   D2
K3  NaN  NaN   C3   D3
"""
# classification of   (left_index, right_on)
res = pd.merge(left, right, left_index=True, how='outer', right_on='D')
print res
"""
      A    B    C   D
K3   A0   B0  NaN  K0
K3   A1   B1  NaN  K1
K3   A2   B2  NaN  K2
K0  NaN  NaN   C0  D0
K2  NaN  NaN   C2  D2
K3  NaN  NaN   C3  D3
"""

  • 首先这里一定要掌握dataframe的column和index的区别否则容易晕
  • 其实left_index和right_index就是和on参数意义一样只是指明使用left dataframe的index和right dataframe的index进行merge
  • index和column也可以一起merge on使用,例如代码中的left_index和right_on

对于重叠(overlapping)


boys = pd.DataFrame({'people': ['name1', 'name2', 'name3'], 'age': [1, 2, 3]})
girls = pd.DataFrame({'people': ['name1', 'name4', 'name5'], 'age': [4, 5, 6]})
print(boys)
print(girls)
"""
   age people
0    1  name1
1    2  name2
2    3  name3
   age people
0    4  name1
1    5  name4
2    6  name5
"""
res = pd.merge(boys, girls, on='people', suffixes=['boys', 'girls'], how='inner')
print(res)
"""
   ageboys people  agegirls
0        1  name1         4
"""

场景为:
统计男生的name和age,女生的name和age,名字的列都用的people,最后两个表格合并的时候,发现有两个重名的同学,所以不能看作一个,作为overlapping的问题解决方案是,
利用suffixes参数进行age的重命名为agegirls, ageboys.

  • 3
    点赞
  • 0
    评论
  • 1
    收藏
  • 一键三连
    一键三连
  • 扫一扫,分享海报

相关推荐
©️2020 CSDN 皮肤主题: Age of Ai 设计师:meimeiellie 返回首页
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、C币套餐、付费专栏及课程。

余额充值