2021-07-27山东大学暑期实训第五周2

---

一、数据处理

import csv
import pymysql

def clean():
    with open("D:/test0727.csv", "r", encoding = "utf-8") as f:
        reader = csv.reader(f)
        rows = [row for row in reader]
    info=[]
    #print(rows)
    for i in range(1,len(rows)):
        #print(rows[i])
        rol=[]
        rol.append(i+39248)
        rol.append("huawei")
        rol.append(int(rows[i][7][:-1]))
        rol.append("无")
        rol.append(rows[i][1])
        if ',' in rows[i][11]:
            if rows[i][11][1]==',':
                price = int(rows[i][11][0])*1000+int(float(rows[i][11][2:]))
            elif rows[i][11][2]==',':
                price = int(rows[i][11][0:2]) * 1000 + int(float(rows[i][11][3:]))
            elif rows[i][11][3]==',':
                price = int(rows[i][11][0:3]) * 1000 + int(float(rows[i][11][4:]))
        elif rows[i][11]=='':
            price=0
        else: price=int(float(rows[i][11]))
        rol.append(price)
        #print(price)
        rol.append(1)
        rol.append("无")
        rol.append(rows[i][6])
        rol.append(int(rows[i][8][:-2]))
        rol.append("2021-07-27")
        info.append(rol)
    return info

info=clean()
print(info[0])

db = pymysql.connect(host="", user="",
                     password="", database="")
cur = db.cursor()
for i in range(len(info)):
    sql = "INSERT INTO huawei_product VALUES ({},'{}',{},'{}','{}',{},{},'{}','{}',{},'{}')"\
        .format(info[i][0],info[i][1],info[i][2],info[i][3],info[i][4],info[i][5],info[i][6],info[i][7],info[i][8],info[i][9],info[i][10])
    try:
       # 执行sql语句
       cur.execute(sql)
       # 提交到数据库执行
       db.commit()
       print("成功",i)
    except:
       # Rollback in case there is any error
       db.rollback()
       print("失败",i)

二、爬取的数据

共有4933条数据