数据分析包Panda学习笔记（二）

3.1 Series.str 包含的常用方法

startwith
endwith
contains
upper
lower
find #找第一个目标字符串出现的下标
split
strip #去空格
replace
slice(0,12)#切片
match#正则匹配
extract#提取正则中的组到每一列

df.City = df.City.str.strip() #去空格，格式化数据

df.City

0            Basildon
1           Parkville
2             Astoria
3              Echuca
4      Cahaba Heights
5           Mickleton
6              Peoria
7              Martin
8            Tel Aviv
9              Chatou
10           New York
11          Eindhoven
12       Shavano Park
13              Eagle
14          Riverside
15        Julianstown
16             Ottawa
17          Hyderabad
18             London
19     Salt Lake City
20         Manchester
21          Ballynora
22         Roodepoort
23        Chula Vista
24             Kuopio
25         Sugar Land
26           New York
27             London
28             Morton
29          Los Gatos
            ...      
968             Perth
969        Greenville
970      Pls Vrds Est
971    Hradec Kralove
972         Maidstone
973      Saint Albans
974         Voluntari
975             Nadur
976          Morrison
977            Cobham
978             Izmir
979         Hollywood
980          Brighton
981            Hawera
982         Worongary
983         Atlantida
984       El Escorial
985          Morrison
986           Augusta
987            Avalon
988            Sydney
989             Lahti
990        Inner City
991             Baden
992      Huddersfield
993           Belfast
994       Black River
995       Skaneateles
996         Haverhill
997           Madison
Name: City, Length: 998, dtype: object

import locale

#将Price去逗号，再转为int形式
df.Price = df.Price.str.replace(',','')

df.Price = df.Price.apply(locate.atoi) #df.apply() 对每一行都运行参数

# order by 排序
df.sort_values('Price',ascending=False)

df.sort_values(['Price','City'],ascending=[False,True]) # 先按price 再按City排 先正序 后倒序

#groupby
df.groupby(df.Country).mean().sort_values('Price',ascending=False) #按平均值分类 再排序

# join on
df2 = pd.DataFrame([['MO',0],['OR',1],['AL',2]],columns=['State','v']) #构建一个新表
df2

pd.merge(df2,df,on='State',how='left')