我试图比较两个表,SQL Server,以验证一些数据。我想从两个表中返回数据在其中一个或另一个中的所有行。本质上,我想展示所有的差异。我需要检查这样做的三段数据,FirstName, LastName和产品。

我对SQL相当陌生,似乎我找到的很多解决方案都过于复杂了。我不需要担心null。

我是这样开始的:

SELECT DISTINCT [First Name], [Last Name], [Product Name] FROM [Temp Test Data]
WHERE ([First Name] NOT IN (SELECT [First Name] 
FROM [Real Data]))

不过我很难继续下去。

谢谢!

编辑:

基于@treaschf的回答,我一直在尝试使用以下查询的变体:

SELECT td.[First Name], td.[Last Name], td.[Product Name]
FROM [Temp Test Data] td FULL OUTER JOIN [Data] AS d 
ON td.[First Name] = d.[First Name] AND td.[Last Name] = d.[Last Name] 
WHERE (d.[First Name] = NULL) AND (d.[Last Name] = NULL)

但是我总是得到0结果,当我知道td中至少有1行不是在d中。

编辑:

好吧,我想我明白了。至少在我几分钟的测试中,它似乎工作得足够好。

SELECT [First Name], [Last Name]
FROM [Temp Test Data] AS td
WHERE (NOT EXISTS
        (SELECT [First Name], [Last Name]
         FROM [Data] AS d
         WHERE ([First Name] = td.[First Name]) OR ([Last Name] = td.[Last Name])))

这基本上会告诉我测试数据中哪些是真实数据中没有的。这完全可以满足我的需求。


当前回答

这将做的技巧,类似于Tiago的解决方案,返回“源”表以及。

select [First name], [Last name], max(_tabloc) as _tabloc
from (
  select [First Name], [Last name], 't1' as _tabloc from table1
  union all
  select [First name], [Last name], 't2' as _tabloc from table2
) v
group by [Fist Name], [Last name]
having count(1)=1

结果将包含表之间的差异,在列_tabloc中,您将有表引用。

其他回答

我知道这可能不是一个受欢迎的答案,但我确实同意@Randy Minder关于在需要更复杂的比较时使用第三方工具的观点。

这里的特定情况很简单,对于这种情况不需要这样的工具,但是如果您引入更多的列、两台服务器上的数据库、更复杂的比较标准等等,情况就很容易变得复杂。

有很多这样的工具,如ApexSQL Data Diff或Quest Toad,你可以在试用模式下使用它们来完成工作。

如果你想要得到哪些列值是不同的,你可以使用实体-属性-值模型:

declare @Data1 xml, @Data2 xml

select @Data1 = 
(
    select * 
    from (select * from Test1 except select * from Test2) as a
    for xml raw('Data')
)

select @Data2 = 
(
    select * 
    from (select * from Test2 except select * from Test1) as a
    for xml raw('Data')
)

;with CTE1 as (
    select
        T.C.value('../@ID', 'bigint') as ID,
        T.C.value('local-name(.)', 'nvarchar(128)') as Name,
        T.C.value('.', 'nvarchar(max)') as Value
    from @Data1.nodes('Data/@*') as T(C)    
), CTE2 as (
    select
        T.C.value('../@ID', 'bigint') as ID,
        T.C.value('local-name(.)', 'nvarchar(128)') as Name,
        T.C.value('.', 'nvarchar(max)') as Value
    from @Data2.nodes('Data/@*') as T(C)     
)
select
    isnull(C1.ID, C2.ID) as ID, isnull(C1.Name, C2.Name) as Name, C1.Value as Value1, C2.Value as Value2
from CTE1 as C1
    full outer join CTE2 as C2 on C2.ID = C1.ID and C2.Name = C1.Name
where
not
(
    C1.Value is null and C2.Value is null or
    C1.Value is not null and C2.Value is not null and C1.Value = C2.Value
)

SQL小提琴示例

@erikkallen回答的简单变化,显示行在哪个表中:

(   SELECT 'table1' as source, * FROM table1
    EXCEPT
    SELECT * FROM table2)  
UNION ALL
(   SELECT 'table2' as source, * FROM table2
    EXCEPT
    SELECT * FROM table1) 

如果你得到一个错误

所有使用UNION、INTERSECT或EXCEPT操作符组合的查询在其目标列表中必须有相等数量的表达式。

然后可能会有帮助

(   SELECT 'table1' as source, * FROM table1
    EXCEPT
    SELECT 'table1' as source, * FROM table2)  
UNION ALL
(   SELECT 'table2' as source, * FROM table2
    EXCEPT
    SELECT 'table2' as source, * FROM table1) 

左连接和大数据的完全连接都存在性能问题。

在我看来,这是最好的解决方案:

select [First Name], count(1) e from
(select * from [Temp Test Data]
 union all
 select * from [Temp Test Data 2]) a
group by [First Name] having e = 1

这将做的技巧,类似于Tiago的解决方案,返回“源”表以及。

select [First name], [Last name], max(_tabloc) as _tabloc
from (
  select [First Name], [Last name], 't1' as _tabloc from table1
  union all
  select [First name], [Last name], 't2' as _tabloc from table2
) v
group by [Fist Name], [Last name]
having count(1)=1

结果将包含表之间的差异,在列_tabloc中,您将有表引用。