【转】用python比对数据库表数据的脚本

最近在做一个数据库异构复制的项目,客户表示需要一个数据比对的工具,我就自己写了一个异构数据库的比对python脚本.这个比对脚本只能比对数量,不能比对具体的记录.使用的sql语句也是最基础的select count(*) 这种,没有开并发所以对大表可能比对时间稍长.

基本原理是将需要比对的数据写到一张表里,先读取那个表里的数据,取出需要比对的表.然后创建多进程,同时在原端和目标端count.然后将count的结果写到一个excel文件中.

其中最关键的就是那张表.只要将那张表里的数据搞对了,基本就不会有什么问题.

目前支持的数据库有oracle,mysql,postgresql,sqlserver.程序分为三个部分

1.数据库配置文件

首先需要在python代码的相同目录下写一个名为check.ini的配置文件.下面一个配置文件例子:

[DATA]
#配置原端数据库,下面的ORACLE需要与后面的项匹配
source=ORACLE
#配置目标端数据库,下面的POSTGRESQL需要与后面的项匹配
target=POSTGRESQL
#配置比对表的数据库,需要与下面的配置项匹配
check_node=ORACLE
#配置比对表数据库的用户,如果是oracle是用户,如果是mysql,pg,mssql则是数据库名
check_owner=suq
#配置比对表的表名,区分大小写
check_table=check_table

#配置mysql的连接串.注意MYSQL必须大写而且必须是以MYSQL开头,例如想比对多个mysql可以写MYSQL1,MYSQL2等
#下面的几个配置同样需要以相应例子开头,因为程序就是以项的开头来确认是哪种数据库的
[MYSQL]
db_host=192.168.56.25
db_port=3306
db_user=root
db_pwd=root
db_dbname=major


[ORACLE]
db_host=192.168.56.30
db_port=1521
db_user=dsg
db_pwd=dsg
db_sid=bre1



[POSTGRESQL]
db_host=192.168.56.50
db_port=5432
db_user=postgres
db_pwd=postgres
db_dbname=msgdb



[MSSQL]
db_host=192.168.56.101
db_port=1433
db_user=sa
db_pwd=sa
db_dbname=master

2.创建一个比对表.

例如我上面的例子放在suq用户下的check_table中

具体的表结构如下:

SQL> desc check_table
 Name                      Null?    Type
 ----------------------------------------- -------- ----------------------------
 SOWNER                         VARCHAR2(30)
 SNAME                          VARCHAR2(30)
 TOWNER                         VARCHAR2(30)
 TNAME                          VARCHAR2(30)

分别表示原端的用户名,表名,目标端用户名表名,如果不是用户的那么就是数据库名.

看一下表内我的测试数据:

SQL> select * from check_table;

SOWNER                 SNAME                  TOWNER                 TNAME
------------------------------ ------------------------------ ------------------------------ ------------------------------
suq                "t1"               suq                t1
suq                "t2"               suq                t2
suq                "t3"               suq                t3
suq                "t4"               suq                t4

这里的数据要特别注意,必须写对否则可能运行会报错.需要注意的一般原因是不同的数据库对大小写敏感不同.因此建议在写好这些数据后,手动到数据库查一下,例如

select count(*) from suq."t1"

看这样的sql对不对.

3.就是主程序

需要注意的是我连接各种数据库分别使用的如下python模块,写excel使用XlsxWriter模块:

C:Users	hink>pip list
cx-Oracle (5.2.1)
MySQL-python (1.2.4)
psycopg2 (2.6.2)
pymssql (2.1.3)
XlsxWriter (0.8.5)

下面是具体的python代码:

#coding:utf-8
import cx_Oracle as ora
import MySQLdb as my
import psycopg2 as post
import pymssql as ms
import ConfigParser as conf
import multiprocessing as  mul
import xlsxwriter 
import time



def connect(cfg,db):
    if db[0:5] == 'MYSQL':
        db_host=cfg.get(db,'db_host')
        db_port=cfg.get(db,'db_port')
        db_user=cfg.get(db,'db_user')
        db_pwd=cfg.get(db,'db_pwd')
        db_dbname=cfg.get(db,'db_dbname')
        conn = my.connect(host=db_host,port=int(db_port),user=db_user,passwd=db_pwd,db=db_dbname)
        return conn
    elif db[0:6] == 'ORACLE':
        db_host=cfg.get(db,'db_host')
        db_port=cfg.get(db,'db_port')
        db_user=cfg.get(db,'db_user')
        db_pwd=cfg.get(db,'db_pwd')
        db_sid=cfg.get(db,'db_sid')
        conn = ora.connect(db_user,db_pwd,db_host+':'+db_port+'/'+db_sid)
        return conn
    elif db[0:10] == 'POSTGRESQL':
        db_host=cfg.get(db,'db_host')
        db_port=cfg.get(db,'db_port')
        db_user=cfg.get(db,'db_user')
        db_pwd=cfg.get(db,'db_pwd')
        db_dbname=cfg.get(db,'db_dbname')
        conn = post.connect(host=db_host,port=db_port,user=db_user,password=db_pwd,database=db_dbname)
        return conn
    elif db[0:5] == 'MSSQL':
        db_host=cfg.get(db,'db_host')
        db_port=cfg.get(db,'db_port')
        db_user=cfg.get(db,'db_user')
        db_pwd=cfg.get(db,'db_pwd')
        db_dbname=cfg.get(db,'db_dbname')
        conn = ms.connect(host=db_host,port=db_port,user=db_user,password=db_pwd,database=db_dbname)
        return conn
        


def check(cfg,db,check_owner,check_table):
    conn=connect(cfg,db)
    cursor=conn.cursor()
    sql='select * from '+check_owner+'.'+check_table
    cursor.execute(sql)
    table_list=[]
    alldata=cursor.fetchall()
    for i in alldata:
        table_list.append(i)
    #print table_list
    return table_list



def getcount(cfg,db,sql,q):
    conn = connect(cfg,db)
    cursor=conn.cursor()
    try:
        cursor.execute(sql)
        countval = cursor.fetchall()[0][0]
        q.put(countval)
    except Exception,e:
        countval="Error : "+str(e)
        q.put(countval)


def isdigit(num):  
    try:  
        int(num)  
        return True  
    except:  
        return False  



def comp(cfg,source,target,tablelist):
    ###excel start
    xlsxname='check_'+str(time.strftime("%Y%m%d%H%M", time.localtime()))+'.xlsx'
    workbook=xlsxwriter.Workbook(xlsxname)
    top=workbook.add_format({'border':6,'align':'center','bg_color':'cccccc','font_size':13,'bold':True})
    format_data_normal=workbook.add_format({'align':'center','font_size':13})
    format_data_warn=workbook.add_format({'align':'center','font_size':13,'bg_color':'ff0000'})
    format_data_err=workbook.add_format({'align':'center','font_size':13,'bg_color':'ffff00'})
    worksheet = workbook.add_worksheet('sheet1')
    worksheet.set_column('A:A',12)
    worksheet.set_column('B:B',40)
    worksheet.set_column('C:C',12)
    worksheet.set_column('D:D',12)
    worksheet.set_column('E:E',40)
    worksheet.set_column('F:F',12)
    worksheet.set_column('G:G',12)
    title=[u'源端用户',u'源端表名',u'源端数据量',u'目标端用户',u'目标端表名',u'目标端数据量',u'差异条数']
    worksheet.write_row('A1',title,top)
    ###excel stop
    length=len(tablelist)
    for i in range(length):
        check_result=[]
        sowner=tablelist[i][0]
        sname=tablelist[i][1]
        towner=tablelist[i][2]
        tname=tablelist[i][3]
        sql_s='select count(*) from '+sowner+'.'+sname
        sql_t='select count(*) from '+towner+'.'+tname
        #sql_t='select count(*) from '+towner+'.'+'"'+tname+'"'
        q1=mul.Queue()
        q2=mul.Queue()
        p1=mul.Process(target = getcount,args = (cfg,source,sql_s,q1))
        p2=mul.Process(target = getcount,args = (cfg,target,sql_t,q2))
        p1.start()
        p2.start()
        count_s=q1.get()
        count_t=q2.get()
        p1.join
        p2.join
        check_result.append(sowner)
        check_result.append(sname)
        check_result.append(count_s)
        check_result.append(towner)
        check_result.append(tname)
        check_result.append(count_t)
        print '%s %s %s %s %s %s'  %(sowner,sname,count_s,towner,tname,count_t)
        #print check_result
        if isdigit(count_s) and isdigit(count_t):
            check_result.append(count_s-count_t)
            if count_s == count_t:
                worksheet.write_row('A'+str(2+i),check_result,format_data_normal)
            else:
                worksheet.write_row('A'+str(2+i),check_result,format_data_warn)
        else:
            check_result.append("Error")
            worksheet.write_row('A'+str(2+i),check_result,format_data_err)
    workbook.close()




    

if __name__ == "__main__":
    print "AT time {0}".format(time.ctime())
    print "Begin compare ..."
    cfg=conf.ConfigParser()
    cfg.read('check.ini')
    source=cfg.get('DATA','source')
    target=cfg.get('DATA','target')
    check_node=cfg.get('DATA','check_node')
    check_owner=cfg.get('DATA','check_owner')
    check_table=cfg.get('DATA','check_table')

    tablelist=check(cfg,check_node,check_owner,check_table)
    comp(cfg,source,target,tablelist)
    print "AT time {0}".format(time.ctime())
    print "compare complete!"
    raw_input("Press <ENTER>")

执行这段代码后就会读取check.ini文件,获取需要比对的原端和目标端数据库的信息,以及比对表的信息,首先将比对的表获取写到一个数组中.然后使用for循环对表进行count,再写到excel中.excel名为check_XXXX.xlsx.xxx为时间.如果在执行sql的时候报错,那么excel中以黄色标出,如果比对原端和目标端数据不一致以红色标出.

下面是我比对oracle和pg中的一个结果:

--------------

# -*- coding:utf-8 -*-
import os

yesterdaynamelist=[]
todaynamelist=[]
differentnamelist=[]
areceivername=[]
test=[]
#读取 昨天生成的namelist 文件 并生成todaynamelist
namelist = open('D:\python\Project\AtuoEmail\Date\riqi.txt','r')
linea = namelist.readlines()
# lineb = namelist.readline()
# print (namelist)
# print (linea)
# print (lineb)
# for i in linea:
#     print (i)
#
for i in linea:
    line=i.split()
    # print(line)
    yesterdaynamelist.extend(line)
    # print(yestdaynamelist)


# 将todaynamelist 列表输出成单列的文本。
yesterdaytxt = open("D:\python\Project\AtuoEmail\Date\yesterdaytxt.txt","w",encoding="utf-8")  #w参数 创建+复写
yesterdaytxt.close()
for i in yesterdaynamelist:
     # print (i)
     # print(type(i))
     yesterdaytxt = open("D:\python\Project\AtuoEmail\Date\yesterdaytxt.txt","a",encoding="utf-8")
     yesterdaytxt.write(i)
     yesterdaytxt.write("
")
yesterdaytxt.close()
# todaytxt.write(todaynamelist)
# todaytxt.close()  #关闭文件

# Yesterdaytxt
# temp

# print (line)
# for i in line:
#     print i.strip().split()[0]
#     print i.strip().split()[1]
#     print i.strip().split()[2]


#调用 cmd生成当天最新的域控用户名单。
os.system('D:\python\Project\AtuoEmail\TodayADUser.bat')

namelist = open('D:\python\Project\AtuoEmail\Date\riqi.txt','r')
linea = namelist.readlines()
todaynamelist=[]
for i in linea:
    line=i.split()
    # print(line)
    todaynamelist.extend(line)
    # print(todaynamelist)

# print(todaynamelist)
# print(yesterdaynamelist)
# print(list(set(todaynamelist).difference(set(yesterdaynamelist))))

#生成差异名单并导出文件
differentnamelist=list(set(todaynamelist).difference(set(yesterdaynamelist))) #“t”“y”对比，输出“T”中新增的元素
# print(differentnamelist)
for i in differentnamelist:
#      print (i)
#      print(type(i))
      differentnamelist = open("D:\python\Project\AtuoEmail\Date\differentnamelist.txt","a",encoding="utf-8")
      differentnamelist.write(i)
      differentnamelist.write("@dafy.com,")
differentnamelist.close()

# -*- coding:utf-8 -*-
import os

yesterdaynamelist=[]
todaynamelist=[]
differentnamelist=[]
areceivername=[]
test=[]
#读取 昨天生成的namelist 文件 并生成todaynamelist
namelist = open('D:\python\Project\AtuoEmail\Date\riqi.txt','r')
linea = namelist.readlines()
# lineb = namelist.readline()
# print (namelist)
# print (linea)
# print (lineb)
# for i in linea:
#     print (i)
#
for i in linea:
    line=i.split()
    # print(line)
    yesterdaynamelist.extend(line)
    # print(yestdaynamelist)


# 将todaynamelist 列表输出成单列的文本。
yesterdaytxt = open("D:\python\Project\AtuoEmail\Date\yesterdaytxt.txt","w",encoding="utf-8")  #w参数 创建+复写
yesterdaytxt.close()
for i in yesterdaynamelist:
     # print (i)
     # print(type(i))
     yesterdaytxt = open("D:\python\Project\AtuoEmail\Date\yesterdaytxt.txt","a",encoding="utf-8")
     yesterdaytxt.write(i)
     yesterdaytxt.write("
")
yesterdaytxt.close()
# todaytxt.write(todaynamelist)
# todaytxt.close()  #关闭文件

# Yesterdaytxt
# temp

# print (line)
# for i in line:
#     print i.strip().split()[0]
#     print i.strip().split()[1]
#     print i.strip().split()[2]


#调用 cmd生成当天最新的域控用户名单。
os.system('D:\python\Project\AtuoEmail\TodayADUser.bat')

namelist = open('D:\python\Project\AtuoEmail\Date\riqi.txt','r')
linea = namelist.readlines()
todaynamelist=[]
for i in linea:
    line=i.split()
    # print(line)
    todaynamelist.extend(line)
    # print(todaynamelist)

# print(todaynamelist)
# print(yesterdaynamelist)
# print(list(set(todaynamelist).difference(set(yesterdaynamelist))))

#生成差异名单并导出文件
differentnamelist=list(set(todaynamelist).difference(set(yesterdaynamelist))) #“t”“y”对比，输出“T”中新增的元素
# print(differentnamelist)
for i in differentnamelist:
#      print (i)
#      print(type(i))
      differentnamelist = open("D:\python\Project\AtuoEmail\Date\differentnamelist.txt","a",encoding="utf-8")
      differentnamelist.write(i)
      differentnamelist.write("@dafy.com,")
differentnamelist.close()