pandas - 行操作

柔情只为你懂 2023-06-08 15:55 80阅读 0赞

import pandas as pd

# operation of line of case #

## sort line by index ##

‘’’\#语法  
df.sort\_index(  
level:specify index by name or Sequence number  
ascending = True:ascending sequence  
inplace = False  
na\_position = ‘last’:(first/last)  
)  
‘’’

df1 = pd.read_csv('123.csv',encoding='utf-8')
    df1.set_index(['地址','店家'],inplace=True)
    
    print(df1.sort_index(level='地址',ascending = False))

## sort line by var value ##

‘’’\# 语法  
df.sort\_values(  
by:specify var name to sort,multi-line  
ascending:True  
inplace = False  
na\_position = ‘last’  
)

print(df1.sort_values(['名次',],ascending = True))

## filter ##

### Essential principle (df灌入True与False序列) ###

sellist = \[True,False,True\]  
print(df1.iloc\[sellist\])

### filter by Absolute position :df.iloc,suitable for row and columns ###

print(df1.iloc[0:3])
    print(df1.iloc[[0,2]])
    print(df1.iloc[:,[0,1,1]])
    print(df1.iloc[:,0:2]) 
    '''range/slice do not need []'''

### filter by index ###

‘’’  
df.loc filter by index ,range filtered include Upper and lower boundary  
will raise error ,when come to unknown index values  
‘’’

df2 = df1.reset_index(['地址','店家'],inplace=False)
    df2.set_index('地址',inplace=True)
    
    print(df2.loc['动物园':'近郊槐阳'])
    
    print(df2.loc[['丰台','昌平'],'店家':'均消费'])

#### Multiple index filter ####

‘’’\#语法  
df.xs(keyword,axis=0,level=None,drop\_level=True)  
specify muliple-level of index  
df.IndexSlice（ fancy usage easy die）

‘’’

df3 = pd.read_csv('123.csv',encoding='utf-8',index_col=[4,3])
    print(df3.loc[[('桥西','无'),('裕华','无')]])
    print(df3.xs('无',level=1,drop_level=False))
    print(df3.xs(['无','桥西'],level=[1,0],drop_level=False))

#### directly Conditional filter ####

‘’’ 1. df3\[df3.columns\_name>10\]’’’  
‘’’ 2. df3.isin(values):\#list filter  
values is sequence:corresponding to every value  
values is dict:corresponding to every variable’s name  
values is df:corresponding to every value and variable’s name in same time’’’

df4 = pd.read_csv('123.csv',encoding='utf-8')
    print(df4[df4.名次<8]) 
    ''' df4.名次<8 形成一个True or False 序列 ，灌回df4'''
    print(df4[df4.名次.isin([3,5,7,9])])
    print(df4[[True for i in range(56) if i%2==0 ]+[False for i in range(56) if not i%2==0 ]]) #筛选的本质
    
    print(df4[[True if i%2==0 else False for i in range(56)]  ])
    print(df4[df4.index.isin([3,5,7,9])])#索引
    print(df4[df4.名次<8][df4[df4.名次<8].名次>5])#主意True与False序列长度必须与DF长度一致

### like SQL method ###

‘’’ df.query(  
expr: like sql expression,use @ cite environment variables  
inplace:False

)  
esample:  
query(’(a < b)&(b < c)’)  
df.query(‘index<b<c’) \# if index do not have a name  
df.query(“code==‘809809’”)  
‘’’

print(df4.query("名次>10 and 名次 < 90 and 地址 in ('桥西','新华')"))
    address  = ['朝阳','大兴']
    print(df4.query("名次>10 and 名次 < 90 and 地址 in @address"))