[DJANGO] excel十几万行数据快速导入数据库研究

时间：2021-07-04 23:50:50

先贴原来的导入数据代码:

8 

import os

os.environ.setdefault("DJANGO_SETTINGS_MODULE", "www.settings") 

'''

Django 版本大于等于1.7的时候，需要加上下面两句

import django

django.setup()

否则会抛出错误 django.core.exceptions.AppRegistryNotReady: Models aren't loaded yet.

'''

import django

if django.VERSION >= (1, 7):#自动判断版本

    django.setup()

from arrears.models import D072Qf

import xlrd #excel读工具

from datetime import datetime

from xlrd import xldate_as_tuple

import time

import random  

time1 = time.time()

#data= xlrd.open_workbook('11.xlsx') 打开文件

with xlrd.open_workbook('11.xlsx') as data:

    print u"读取文件结束,开始导入!"

    time2 = time.time()

    table = data.sheet_by_index(0) #获取工作表

    time3 = time.time()

    n=1

    x = y = z = 0

    WorkList = []

    for line in range(n,table.nrows):#nrows = table.nrows #行数 ncols = table.ncols #列数 print sh.row_values(rownum)

        row = table.row_values(line)

        if row: #查看行值是否为空

            for i in [0,1,2,4,28,30,32]:

                if type(row[i]) == float:

                    row[i] = int(row[i])

            if D072Qf.objects.filter(acct_month = row[0],serv_id=row[1]).exists():#判断该行值是否在数据库中重复

                x = x + 1 #重复值计数

            else:

                WorkList.append(D072Qf(acct_month=row[0],serv_id=row[1],acc_nbr=row[2],user_name=row[3],acct_code=row[4],

                                   acct_name=row[5],product_name=row[6],current_charge=row[7],one_charge=row[8],

                                   two_charge=row[9],three_charge=row[10],four_charge=row[11],five_charge=row[12],

                                   six_charge=row[13],seven_charge=row[14],eight_charge=row[15],nine_charge=row[16],

                                   ten_charge=row[17],eleven_charge=row[18],twelve_charge=row[19],oneyear_charge=row[20],

                                   threeyear_charge=row[21],upthreeyear_charge=row[22],all_qf=row[23],morethree_qf=row[24],

                                   aging=row[25],serv_state_name=row[26],mkt_chnl_name=row[27],mkt_chnl_id=row[28],

                                   mkt_region_name=row[29],mkt_region_id=row[30],mkt_grid_name=row[31],mkt_grid_id=row[32],

                                   prod_addr=row[33]))

                y = y + 1 #非重复计数

        else:

             z = z + 1     #空行值计数

        n = n + 1

        if n % 9999 == 0:

            D072Qf.objects.bulk_create(WorkList)

            WorkList = []

            time.sleep(random.random()) #让Cpu随机休息0 <= n < 1.0 s

            print "导入成功一次!"

            print '数据导入成功,导入'+str(y)+'条,重复'+str(x)+'条,有'+str(z)+'行为空!'

            time4 = time.time()

            print "读取文件耗时"+str(time2-time1)+"秒,导入数据耗时"+str(time4-time3)+"秒!"

这条代码目前未全部将十几万行数据全部导入数据库中,只花了1个小时把5万行数据导入其中后,后面越来越慢,主要慢在excel表到了7万行数据左右后,读取excel中数据很慢了,总体来说影响导入速度有几个原因:

1、一直以来采用xlrd导入xls格式文件,如果文件有十几万行,只是读取文件就会花200秒,若换成csv则几乎不花时间

2、代码中这行语句也会影响速度,特别当数据库中数据很大时:if D072Qf.objects.filter(acct_month = row[0],serv_id=row[1]).exists():#判断该行值是否在数据库中重复

3、若一次性将字典添加十几万行数据,就windows的cpu而已是遭受不住的!所以建议1万条数据导入一次后,清空列表

改善后的代码:

优化部分:采用csv格式;取消掉检查重复数据语句;每5万导入一次数据

#coding:utf-8 

import os

os.environ.setdefault("DJANGO_SETTINGS_MODULE", "www.settings") 

'''

Django 版本大于等于1.7的时候，需要加上下面两句

import django

django.setup()

否则会抛出错误 django.core.exceptions.AppRegistryNotReady: Models aren't loaded yet.

'''

import django

if django.VERSION >= (1, 7):#自动判断版本

    django.setup()

from arrears.models import D072Qf

import time

import random

time1 = time.time()

f = open('11.csv')

print u"读取文件结束,开始导入!"

time2 = time.time()

WorkList = []

next(f) #将文件标记移到下一行

y = 0

n = 1

for line in f:

    row = line.replace('"','') #将字典中的"替换空

    row = row.split(';') #按;对字符串进行切片

    y = y + 1

    WorkList.append(D072Qf(acct_month=row[0],serv_id=row[1],acc_nbr=row[2],user_name=row[3],acct_code=row[4],

                                   acct_name=row[5],product_name=row[6],current_charge=row[7],one_charge=row[8],

                                   two_charge=row[9],three_charge=row[10],four_charge=row[11],five_charge=row[12],

                                   six_charge=row[13],seven_charge=row[14],eight_charge=row[15],nine_charge=row[16],

                                   ten_charge=row[17],eleven_charge=row[18],twelve_charge=row[19],oneyear_charge=row[20],

                                   threeyear_charge=row[21],upthreeyear_charge=row[22],all_qf=row[23],morethree_qf=row[24],

                                   aging=row[25],serv_state_name=row[26],mkt_chnl_name=row[27],mkt_chnl_id=row[28],

                                   mkt_region_name=row[29],mkt_region_id=row[30],mkt_grid_name=row[31],mkt_grid_id=row[32],

                                   prod_addr=row[33]))

    n = n + 1

    if n%50000==0:

        print n

        D072Qf.objects.bulk_create(WorkList)

        WorkList = []

        time3 = time.time()

        print "读取文件耗时"+str(time2-time1)+"秒,导入数据耗时"+str(time3-time2)+"秒!"

time3 = time.time()

print n

D072Qf.objects.bulk_create(WorkList)

print "读取文件耗时"+str(time2-time1)+"秒,导入数据耗时"+str(time3-time2)+"秒!"

WorkList = []

print "成功导入数据"+str(y)+"条"

f.close()

结果让人大吃一惊!!!,只耗时73秒

Python 2.7.10 (default, May 23 2015, 09:40:32) [MSC v.1500 32 bit (Intel)] on win32

Type "copyright", "credits" or "license()" for more information.

>>> ================================ RESTART ================================

>>>

读取文件结束,开始导入!

50000

读取文件耗时0.0秒,导入数据耗时34.3279998302秒!

100000

读取文件耗时0.0秒,导入数据耗时67.3599998951秒!

138400

读取文件耗时0.0秒,导入数据耗时73.4379999638秒!

成功导入数据138399条

>>>

相关文章