Python_PyMySQL数据库操作

连接数据库:

conn=pymysql.connect(host='*.*.*.*',port=3306,user='root',password='123456',charset='utf8')

建立游标:

cur = conn.cursor()

创建一个名字叫 lj 的数据库:

cur.execute('create database if not EXISTS lj character set utf8;') #用cur.execute执行SQL语句　　注意如果要插入中文数据,建库时必须要加 character set utf8 意思是把字符集设置为utf8,要不然中文不能显示!

选则要操作的数据库:

conn.select_db('lj')

添加7个字段:

cur.execute('create table JiNan(source VARCHAR(50),addr VARCHAR(50),stru VARCHAR(50),dir_1 VARCHAR(50),floor VARCHAR(50),size VARCHAR(50),money VARCHAR(50));')

提交修改:

conn.commit()

向表中插入数据:

cur.execute("insert into JiNan(source,addr,stru,dir_1,floor,size,money) VALUES('%s','%s','%s','%s','%s','%s','%s')"%(so,ad,st,di,fl,si,mn))

整个爬取网站数据然后存入数据库的脚本:

#-*- coding:utf-8 -*-
import requests,re,pymysql
def get_info(url):
    html_response = requests.get(url)
    html_response.encoding = 'utf-8'
    add_add = r'<span class="region">(.*?)&nbsp;&nbsp;'
    stru_add = r'class="zone"><span>(.*?)&nbsp;&nbsp;'
    size_add = r'<span class="meters">(.*?)&nbsp;&nbsp;'
    dir_add = r'</span><span>(.*?)</span></div><div'
    floor_add = r'<span>/</span>(.*?)</div></div><div'
    source_add = r'<div class="con"><a href=".*?">(.*?)</a>'
    money_add = r'lass="price"><span class="num">(.*?)</span>元/月'

    add    = re.findall(add_add,html_response.text)
    stru   = re.findall(stru_add,html_response.text)
    size   = re.findall(size_add,html_response.text)
    dir_1  = re.findall(dir_add,html_response.text) 
    floor  = re.findall(floor_add,html_response.text)
    source = re.findall(source_add,html_response.text)
    money  = re.findall(money_add,html_response.text)

    #print(add[0],stru[0],dir_1[0],floor[0],source[0],size[0],money[0])
    return source,add,stru,dir_1,floor,size,money
def into_db(source,add,stru,dir_1,floor,size,money,cur,conn):
    for so,ad,st,di,fl,si,mn in zip(source,add,stru,dir_1,floor,size,money):
        cur.execute("insert into JiNan(source,addr,stru,dir_1,floor,size,money) VALUES('%s','%s','%s','%s','%s','%s','%s')"%(so,ad,st,di,fl,si,mn))
        conn.commit()


if __name__ == '__main__':
    
    conn = pymysql.connect(host='*.*.*.*',port=3306,user='root',password='1234',charset='utf8')
    cur = conn.cursor()
    cur.execute('create database if not EXISTS lj character set utf8;')
    conn.select_db('lj')
    cur.execute('create table JiNan(source VARCHAR(50),addr VARCHAR(50),stru VARCHAR(50),dir_1 VARCHAR(50),floor VARCHAR(50),size VARCHAR(50),money VARCHAR(50));')
    conn.commit()
    for i in range(1,101):
        url = 'https://jn.lianjia.com/zufang/pg' + str(i)
        print(url)
        source,add,stru,dir_1,floor,size,money = get_info(url)
        into_db(source,add,stru,dir_1,floor,size,money,cur,conn)
        print('OK... ...')
    conn.close()

View Code