[DJANGO] excel十幾萬行數據快速導入數據庫研究

先貼原來的導入數據代碼:

8 

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

'''
Django 版本大於等於1.7的時候，需要加上下面兩句
import django
django.setup()
否則會拋出錯誤 django.core.exceptions.AppRegistryNotReady: Models aren't loaded yet.
'''
import django

if django.VERSION >= (1, 7):#自動判斷版本
    django.setup()

from arrears.models import D072Qf
import xlrd #excel讀工具
from datetime import datetime
from xlrd import xldate_as_tuple
import time
import random  

time1 = time.time()
#data= xlrd.open_workbook('11.xlsx') 打開文件
with xlrd.open_workbook('11.xlsx') as data:
    print u"讀取文件結束,開始導入!"
    time2 = time.time()
    table = data.sheet_by_index(0) #獲取工作表
    time3 = time.time()
    n=1
    x = y = z = 0
    WorkList = []
    for line in range(n,table.nrows):#nrows = table.nrows #行數 ncols = table.ncols #列數 print sh.row_values(rownum)
        row = table.row_values(line)
        if row: #查看行值是否為空
            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():#判斷該行值是否在數據庫中重復
                x = x + 1 #重復值計數
            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 #非重復計數                
        else:
             z = z + 1     #空行值計數
        n = n + 1
        if n % 9999 == 0:
            D072Qf.objects.bulk_create(WorkList)                     
            WorkList = []
            time.sleep(random.random()) #讓Cpu隨機休息0 <= n < 1.0 s
            print "導入成功一次!"
            print '數據導入成功,導入'+str(y)+'條,重復'+str(x)+'條,有'+str(z)+'行為空!'     
            time4 = time.time()
            print "讀取文件耗時"+str(time2-time1)+"秒,導入數據耗時"+str(time4-time3)+"秒!"

這條代碼目前未全部將十幾萬行數據全部導入數據庫中,只花了1個小時把5萬行數據導入其中后,后面越來越慢,主要慢在excel表到了7萬行數據左右后,讀取excel中數據很慢了,總體來說影響導入速度有幾個原因:

1、一直以來采用xlrd導入xls格式文件,如果文件有十幾萬行,只是讀取文件就會花200秒,若換成csv則幾乎不花時間

2、代碼中這行語句也會影響速度,特別當數據庫中數據很大時:if D072Qf.objects.filter(acct_month = row[0],serv_id=row[1]).exists():#判斷該行值是否在數據庫中重復

3、若一次性將字典添加十幾萬行數據,就windows的cpu而已是遭受不住的!所以建議1萬條數據導入一次后,清空列表

改善后的代碼:

優化部分:采用csv格式;取消掉檢查重復數據語句;每5萬導入一次數據

#coding:utf-8 

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

'''
Django 版本大於等於1.7的時候，需要加上下面兩句
import django
django.setup()
否則會拋出錯誤 django.core.exceptions.AppRegistryNotReady: Models aren't loaded yet.
'''
import django

if django.VERSION >= (1, 7):#自動判斷版本
    django.setup()

from arrears.models import D072Qf 
import time
import random
time1 = time.time()
f = open('11.csv')
print u"讀取文件結束,開始導入!"
time2 = time.time()
WorkList = []
next(f) #將文件標記移到下一行
y = 0
n = 1
for line in f:
    row = line.replace('"','') #將字典中的"替換空
    row = row.split(';') #按;對字符串進行切片
    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 "讀取文件耗時"+str(time2-time1)+"秒,導入數據耗時"+str(time3-time2)+"秒!"
time3 = time.time()
print n
D072Qf.objects.bulk_create(WorkList)
print "讀取文件耗時"+str(time2-time1)+"秒,導入數據耗時"+str(time3-time2)+"秒!"
WorkList = []
print "成功導入數據"+str(y)+"條"
f.close() 

結果讓人大吃一驚!!!,只耗時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 ================================
>>> 
讀取文件結束,開始導入!
50000
讀取文件耗時0.0秒,導入數據耗時34.3279998302秒!
100000
讀取文件耗時0.0秒,導入數據耗時67.3599998951秒!
138400
讀取文件耗時0.0秒,導入數據耗時73.4379999638秒!
成功導入數據138399條
>>>