python 爬虫 全国各地车牌号爬取自编

运行效果：


代码实现：

#!/usr/bin/env python
# -*- coding: utf-8 -*-
# @Time    : 2020/11/24 8:51
# @Author  : huni
# @File    : 车牌号.py
# @Software: PyCharm

import requests
from lxml import etree
import sqlite3

def getData(url):
    headers = {
         'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/86.0.4240.183 Safari/537.36'}

    cata = requests.get(url=url,headers=headers).text

    # # 处理中文乱码问题
    # cata = cata.text.encode('ISO-8859-1').decode('utf-8')

    tree = etree.HTML(cata)

    mcon_list = tree.xpath('//div[@class="mcon"]//text()')[81:]

    # print(mcon_list)
    all_info_list = []
    for mcon in mcon_list:
        pai = []
        pai.append(mcon[:4].strip('【】'))
        pai.append(mcon[4:])
        all_info_list.append(pai)
    return all_info_list

#数据保存到数据库
def saveDatadb(dbpath,all_info_list):
    init_db(dbpath)
    conn = sqlite3.connect(dbpath)
    cur = conn.cursor()

    for data in all_info_list:
        for index in range(len(data)):
            data[index] = '"'+str(data[index])+'"'      #replace into
        sql = '''
                insert into 车牌号表(
                车牌前两位,所属地区)
                values(%s)'''%",".join(data)
        # print(sql)
        cur.execute(sql)
        conn.commit()
    cur.close()
    conn.close()



#初始化数据库，创建表
def init_db(dbpath):
    sql1 = '''
            drop table if exists 车牌号表
        '''  # 删除原来的数据表
    sql2 = '''
        create table if not exists 车牌号表
        (
        车牌前两位 varchar ,
        所属地区 varchar 
        )

    '''  # 创建数据表
    conn = sqlite3.connect(dbpath)
    cursor = conn.cursor()
    cursor.execute(sql1)
    cursor.execute(sql2)
    conn.commit()
    conn.close()


def main():
    url = 'https://chepai.911cha.com/'
    all_info_list = getData(url)
    dbpath = '车牌号.db'
    saveDatadb(dbpath,all_info_list)

if __name__ == '__main__':
    main()
    print('保存成功')