pandas 变量列的基本操作

客官°小女子只卖身不卖艺 2023-06-06 08:28 17阅读 0赞

import pandas as pd
    
    df1 = pd.read_csv('123.csv',encoding = 'utf-8')

# simply review data 查看数据 #

print(df1)# directly show dataframe 直接查看
    print(df1.info)#dataframe's basic information 信息描述
    print(df1.head(10))#glance over top 10 records 看前十行
    print(df1.tail(10))#glance over last 10 records 看后十行

# rename columns’ name 重命名列名 #

print(df1.columns) #show columns' name 打印所有列名称
    df1.columns = ['storename','star','consumption','address','phone']
    print(df1.columns)
    df2 = df1.rename(
            columns = { 'star':'stars'},#note wrong name will be ignored 错误或者没有的列名将忽略不报错
            inplace = False #if or not directly replace dataframe
            )
    print(df2)

# filter columns 筛选列 #

df.var: Applicable to existing columns, only select single column
    df['var'] single columns outcome is series,use list can change it to df
    df[['var1','var2']] multi-columns' usage,var not exit

# 删除列 #

del df['columns-name'] # directly delet df columns
    print(df2[['stars','phone']]) #delet columns 
    
    df.drop( 
            index/columns = row/columns' label ready to delet ,more than one use list
            inplace = False
    )
    #删除行和列
    print(df2.drop(columns = ['stars','phone']))
    print(df2.drop(index = [0,1]))

# 数据类型操作 #

pandas' data type:#所支持的类型
        float,int,string,bool,datetime64[ns],datetime64[ns,tz],timedelta[ns]
        category,object
    print(df2.dtypes)

# convert data types #

df2.astype(
            dtype:assign new type you want (int/float/bool/str)
            copy:same to inplace; True is default 
            erroes = 'raise':if or not raise errors [ignore/raise]     
        )
    '''
    print(df2.astype('str').dtypes)#所有列转换为str
    print(df2.stars.astype('float'))#列stars转换为float
    print(df2.stars.astype('float').dtypes)
    print(df2[['stars','phone']].astype('str').dtypes)列stars和phone转换为str
    print(df2[['stars','phone']].astype('str').apply(pd.to_numeric,errors = 'ignore').dtypes)
     # apply :Batch processing cmd to subset 批处理命令，对子集操作
    
    df3 = pd.DataFrame({ 'A':['a',1,2,3]})
    df = df3.iloc[1:]
    print(df )
    print(df.dtypes)
    print(df3.infer_objects().dtypes)#auto-convert types 自动识别类型并转换，经测试并不好用