Select 与 sql table 不同的当前列和先前列

Select distinct current and previous columns from a sql table

我有一个table像这样

   Id  |Name    |Status |   Rate |  Method  |ModifiedTime             |ModifiedBy
-----------------------------------------------------------------------------
    1  |Recipe1 |  0    |    30  |   xyz    | 2016-07-26 14:55:57.977 |     A
-------------------------------------------------------------------------------
    2  |Recipe1 |  0    |    30  |   abc    | 2016-07-26 14:56:18.123 |     A
--------------------------------------------------------------------------------
    3  |Recipe1 |  1    |    30  |   xyz    | 2016-07-26 14:57:50.180 |     b

我只想 select 更改,并希望显示更改值之前的值和当前值以及更改者的值。最终结果如下。我正在使用 SQL Server 2014。

Item    | Before |  After |ModifiedTime             | ModifiedBy
-----------------------------------------------------------------------------
Method  |  xyz   |  Abc   | 2016-07-26 14:56:18.123 |   A
-------------------------------------------------------------------------------
Status  |  0     |  1     | 2016-07-26 14:57:50.180 |   b
--------------------------------------------------------------------------------
Method  |  Abc   |  xyz   | 2016-07-26 14:57:50.180 |   b

没有动态 SQL 这是我能做的最好的。 我假设您按 Name 对更改进行分组。如果 Name 可以更改,则应将其从分区中删除并添加到项目子查询和 case 中。你可以试试here

select item, 
       case item
         when 'Status' then cast(prevStatus as varchar)
         when 'Rate' then cast(prevRate as varchar)
         when 'Method' then prevMethod
         end as Before, 
       case item
         when 'Status' then cast(Status as varchar)
         when 'Rate' then cast(Rate as varchar)
         when 'Method' then Method
         end as After,
         ModifiedTime,
         ModifiedBy
  from (
    select Status,
           lag(Status) over (partition by Name order by id) prevStatus,
           Rate,
           lag(Rate) over (partition by Name order by id) prevRate,
           Method,
           lag(Method) over (partition by Name order by id) prevMethod,
           ModifiedBy,
           ModifiedTime
      from t ) as t1 cross join  (select 'Status' as item union all 
                                  select 'Rate' as item union all 
                                  select 'Method' as item) items
where (item = 'Status' and Status <> prevStatus)
   or (item = 'Rate' and Rate <> prevRate)
   or (item = 'Method' and Method <> prevMethod)
order by ModifiedTime

输出

item   Before After ModifiedTime            ModifiedBy 
------ ------ ----- ----------------------- ---------- 
Method xyz    abc   2016-07-26 14:56:18.123 A          
Status 0      1     2016-07-26 14:57:50.180 b          
Method abc    xyz   2016-07-26 14:57:50.180 b   

   

在第二行你会看到 ModifiedBy 是 b 而你有 A。我认为是错别字。