Question

假设我在SQL Server DB中有这个表：

Date     Col_A    Col_B    Col_C    ...
Day1     1        2        3    ....
Day1     2        3        4    ....
Day1     3        4        5    ....
Day2     1        2        3    ....
Day2     2        3        4    ....
Day2     5        6        7    ....
Day2     6        7        8    ....

我希望根据Col_A，Col_B＆amp; Col_C获得如下内容：

Date     Col_A    Col_B    Col_C    ...  Date   Col_A    Col_B    Col_C   ....
Day1     1        2        3    ....     Day2     1        2        3    ....
Day1     2        3        4    ....     Day2     2        3        4    ....
Day1     3        4        5    ....     NULL    NULL    NULL     NULL  ..NULL..
NULL    NULL    NULL     NULL ..NULL..   Day2     5        6        7    ....
NULL    NULL    NULL     NULL ..NULL..   Day2     6        7        8    ....

我认为这个SQL会这样做：

SELECT * FROM
   (Select * from MyTable where Date = 'Day1') T1
FULL OUTER JOIN 
   (Select * from MyTable where Date = 'Day2') T2
ON 
   T1.Col_A = T2.Col_A
  AND
   T1.Col_B = T2.Col_B
  AND
   T1.Col_C = T2.Col_C

但这只是返回T1已满且T2有NULLS的行，而不是完整的连接。

我做错了什么，如何解决这个问题？另外，是否有更好的方法来获得这种记录集，而不是将表连接到自身并只更改日期参数??

感谢!!!

Answer 1

你的加入看起来不错，请尝试这样的事情以确保（正确命名所有列）：

select
    isnull(T1.Col_A, T2.Col_A) as Col_A,
    isnull(T1.Col_B, T2.Col_B) as Col_B,
    isnull(T1.Col_A, T2.Col_C) as Col_C,
    T1.Date as Date1,
    T2.Date as Date2
from (select * from MyTable where Date = 'Day1') as T1
    full outer join (select * from MyTable where Date = 'Day2') as T2 on 
       T1.Col_A = T2.Col_A and T1.Col_B = T2.Col_B and T1.Col_C = T2.Col_C
order by 1, 2, 3

为什么这个全连接不起作用？

1 个答案: