目录

  • merge()函数介绍
    • 说明
    • 参数说明:
    • 参考文件:

merge()函数介绍

说明

pandas.merge(left, right, how: str = 'inner', on=None, left_on=None, right_on=None, left_index: bool = False, right_index: bool = False, sort: bool = False, suffixes='_x', '_y', copy: bool = True, indicator: bool = False, validate=None)

功能:用于合并两个 DataFrame 对象或 Series对象。只能用于两个表的拼接(左右拼接,不能用于上下拼接)
应用场景数据合并 ( 数据合并的另一个常用函数是pd.concat()

参数说明:

left,right
用于拼接的两个表中,即使没有确定谁是主键,函数也会自动将两个表中的重复列作为主键,直接把一个表的名字传递给参数left,另一个表的名字传递给参数right

how
参数拼接方式,默认内连接(‘inner’)
内连接inner:将两个表主键一致的信息拼接到一起
外连接outer:保留两个表的所有信息,如果遇到对不齐的部分,用NAN填充
左连接left:保留左表的全部信息,把右表满足主键的行信息并进来,对不齐的部分用NAN填充
右连接right:保留右表的全部信息(同上)

示例:

import pandas as pddata1 =pd.DataFrame({'a':['a1','a2','a3'], 'b':['b1','b2','b3'], 'key':['a','b','c'], 'key1':['d','e','f']})data2 = pd.DataFrame({'c':['c1','c2','c3'],'d':['d1','d2','d3'],'key':['a','b','a'],'key1':['d','e','e']})result=pd.merge(data1,data2,on = ['key','key1'])result1=pd.merge(data1,data2,how ="left",on = ['key','key1'])result2=pd.merge(data1,data2,how ="right",on = ['key','key1'])result3=pd.merge(data1,data2,how ="inner",on = ['key','key1'])result4=pd.merge(data1,data2,how ="outer",on = ['key','key1'])print(data1)print(data2)print(result)print(result1)print(result2)print(result3)print(result4)

运行结果:

a b key key10a1b1 ad1a2b2 be2a3b3 cfc d key key10c1d1 ad1c2d2 be2c3d3 aea b key key1 c d0a1b1 adc1d11a2b2 bec2d2a b key key1cd0a1b1 ad c1 d11a2b2 be c2 d22a3b3 cfNaNNaN ab key key1 c d0 a1 b1 adc1d11 a2 b2 bec2d22NaNNaN aec3d3a b key key1 c d0a1b1 adc1d11a2b2 bec2d2 ab key key1cd0 a1 b1 ad c1 d11 a2 b2 be c2 d22 a3 b3 cfNaNNaN3NaNNaN ae c3 d3Process finished with exit code 0

on
1、确定哪个字段作为主键
2、如果两个表中有两列以上信息相同,可以指定哪一列作为主键,如果不指定,相同信息的列都会作为拼接依据
3、merge()函数默认的是内连接,因此只拼接两表中拥有相同主键信息的行数据。

示例:

import pandas as pddata1 =pd.DataFrame({'key':['K0','K1','K2','K3'],'A':['A0','A1','A2','A3'],'B':['B0','B1','B2','B3']})data3= pd.DataFrame({'key':['K0','K1','K2','K3'],})data2 = pd.DataFrame({'key':['K0','K1','K2','K3'],'C':['C0','C1','C2','C3'],'D':['D0','D1','D2','D3']})result1 = pd.merge(data1,data3,on = 'key')result2 = pd.merge(data1,data2,on = 'key')result3 = pd.merge(data3,data2,on = 'key')print(data1)print(data2)print(data3)print(result1)print(result2)print(result3)

运行结果:

 key A B0K0A0B01K1A1B12K2A2B23K3A3B3key C D0K0C0D01K1C1D12K2C2D23K3C3D3key0K01K12K23K3key A B0K0A0B01K1A1B12K2A2B23K3A3B3key A B C D0K0A0B0C0D01K1A1B1C1D12K2A2B2C2D23K3A3B3C3D3key C D0K0C0D01K1C1D12K2C2D23K3C3D3Process finished with exit code 0

left-index, right-index
1、除指定字段可以作为主键外,索引也可以考虑作为拼接的主键
2、默认为False,即不以索引为主键
3、如果两个表的索引完全一样,直接拼接效果很好,如果索引有不能对齐的地方,在默认的内连接情况下,只会把索引对齐的记录进行拼接

示例:

import pandas as pddf1 = pd.DataFrame({'lkey': ['foo', 'ba', 'baz', 'fo'],'value': [1, 2, 3, 4]},index=['A', 'B', 'C', 'D'])df2 = pd.DataFrame({'rkey': ['foo', 'bar', 'baz', 'foo'], 'value': [5, 6, 7, 8]},index=['A', 'c', 'B', 'h'])# result1 = pd.merge(df1,df2,left_on ='lkey')result2 = pd.merge(df1,df2,left_on ='lkey',right_on ='rkey')result3 = pd.merge(df1,df2,left_on ='lkey',right_on ='rkey',suffixes=("_lf","_rf"))result4=pd.merge(df1,df2,left_index=True, right_index=True)#result6=pd.merge(df1,df2,left_index=True)result5=pd.merge(df1,df2,left_index=True, right_index=True,suffixes=("_lf","_rf"))print(df1)print(df2)# print(result1)print(result2)print(result3)print(result4)print(result5)#print(result6)

运行结果:

lkeyvalueAfoo1B ba2Cbaz3D fo4rkeyvalueAfoo5cbar6Bbaz7hfoo8lkeyvalue_x rkeyvalue_y0foo1foo51foo1foo82baz3baz7lkeyvalue_lf rkeyvalue_rf0foo 1foo 51foo 1foo 82baz 3baz 7lkeyvalue_x rkeyvalue_yAfoo1foo5B ba2baz7lkeyvalue_lf rkeyvalue_rfAfoo 1foo 5B ba 2baz 7Process finished with exit code 0

left_on,right_on
两个表里没有完全一致的列名,但是有信息一致的列
该参数用来指定用来作主键的列名是哪一个
需要保证键值长度相等,len(left_on) == len(right_on)

suffixes
两个表中出现相同的列名,除了作为主键的列之外,其他名字相同的列被拼接到表中的时候会有一个后缀表示这个列来自于哪个表格,用于区分名字相同的列,这个后缀默认是(x和y)。这个后缀是可以自定义修改的

示例:

import pandas as pddf1 = pd.DataFrame({'lkey': ['foo', 'ba', 'baz', 'fo'],'value': [1, 2, 3, 4]})df2 = pd.DataFrame({'rkey': ['foo', 'bar', 'baz', 'foo'], 'value': [5, 6, 7, 8]})# result1 = pd.merge(df1,df2,left_on ='lkey')result2 = pd.merge(df1,df2,left_on ='lkey',right_on ='rkey')result3 = pd.merge(df1,df2,left_on ='lkey',right_on ='rkey',suffixes=("_lf","_rf"))print(df1)print(df2)# print(result1)print(result2)print(result3)

运行结果:

lkeyvalue0foo11 ba22baz33 fo4rkeyvalue0foo51bar62baz73foo8lkeyvalue_x rkeyvalue_y0foo1foo51foo1foo82baz3baz7lkeyvalue_lf rkeyvalue_rf0foo 1foo 51foo 1foo 82baz 3baz 7Process finished with exit code 0

indicator
用于显示拼接后的表中信息来自哪个表
在表的最后一列显示left_only /right_only/both
默认False,可以修改为True
示例:

import pandas as pddata1 =pd.DataFrame({'key':['K','1','K2','K3'], 'A':['A0','1','A2','A3'], 'B':['B0','B1','B2','B3']})data2 = pd.DataFrame({'key':['K0','K1','K2','K3'],'C':['C0','C1','C2','C3'],'D':['D0','D1','D2','D3']})result=pd.merge(data1,data2,how ='outer',on = ['key'],indicator = True)print(data1)print(data2)print(result)

运行结果:

key A B0 KA0B01 1 1B12K2A2B23K3A3B3key C D0K0C0D01K1C1D12K2C2D23K3C3D3keyABCD_merge0 K A0 B0NaNNaN left_only1 11 B1NaNNaN left_only2K2 A2 B2 C2 D2both3K3 A3 B3 C3 D3both4K0NaNNaN C0 D0right_only5K1NaNNaN C1 D1right_onlyProcess finished with exit code 0

参考文件:

https://zhuanlan.zhihu.com/p/340770510