比较来自不同数据库的记录时结果不正确

Question

我有以下数据

我想比较这些数据。所以，我使用这段代码：

drop view sms4_table;
create view sms4_table as
select snumber,cast(to_char(cdrdate,'YYYYMMDDHH24MISS')as varchar2(14))as cdrdate
      from database1@db
      where snumber='123456789101098'
and (cdrdate between to_date ('05/07/2014 00:00:00','DD/MM/YYYY HH24:MI:SS')
and to_date ('05/07/2014 23:59:59','DD/MM/YYYY HH24:MI:SS'));

drop view sms3_table;
create view sms3_table as
select cast (substr(snumber,1,15) as varchar2(15)) as snumber,timestamp as answertime
     FROM database2 
          where snumber='123456789101098A';
                and timestamp like '20140705%'

select sms4_table.snumber, sms4_table.cdrdate, sms3_table.snumber, sms3_table.answertime
       from sms4_table
       left join sms3_table
on  sms4_table.cdrdate=sms3_table.answertime
and sms4_table.snumber=sms3_table.snumber
where sms3_table.snumber is null;

我在database1中有 20 条记录，在database2中有 19 条记录。数据库之间的唯一区别是数据库1中的两个记录是重复的（这与snumber 123456789101098和cdrdate 20140705033021）。但是，当我运行上面的脚本时，我没有收到差异。我只收到 -

  

选择了0行

请你支持我。如何纠正脚本以显示这种差异 - snumber 123456789101098和cdrdate 20140705033021？

感谢您提前帮助我。

Answer 1

您可以向两个视图添加行号伪列，以便可以区分重复值：

create view sms4_table as
select snumber,
  to_char(cdrdate,'YYYYMMDDHH24MISS') as cdrdate,
  row_number() over (partition by snumber,
    to_char(cdrdate,'YYYYMMDDHH24MISS')) as rn
from database1@db
where snumber='123456789101098'
and cdrdate between to_date ('05/07/2014 00:00:00','DD/MM/YYYY HH24:MI:SS')
and to_date ('05/07/2014 23:59:59','DD/MM/YYYY HH24:MI:SS');

create view sms3_table as
select substr(snumber,1,15) as snumber,
  timestamp as answertime,
  row_number() over (partition by substr(snumber,1,15), timestamp) as rn
from database2 
where snumber='123456789101098A'
and timestamp like '20140705%';

然后在连接中包含rn伪列;我已将其更改为完全外部联接，因此您可以以任何方式发现问题。

select sms4_table.snumber, sms4_table.cdrdate,
  sms3_table.snumber, sms3_table.answertime
from sms4_table
full outer join sms3_table
on sms4_table.cdrdate = sms3_table.answertime
and sms4_table.snumber = sms3_table.snumber
and sms4_table.rn = sms3_table.rn
where sms3_table.snumber is null
or sms4_table.snumber is null;

对于您的重复值sms4_table现在将具有：

snumber         cdrdate         rn
--------------- -------------- ---
123456789101098 20140705033021   1
123456789101098 20140705033021   2

而sms3_table只有：

snumber         answertime      rn
--------------- -------------- ---
123456789101098 20140705033021   1

如果您只加入前两列，那么视图4中的两个记录都将匹配视图3中的记录 - 正如您的原始查询所做的那样。但是，如果在连接条件中包含rn，则只有第一行匹配，而rn = 2的第二行不匹配。

如果这是一次性任务，你真的不需要创建视图，你可以只使用内联视图（子查询）。