如何在没有任何存储过程或函数的情况下删除SQL中的中间值？

Question

我正在处理一些报告生成脚本，在这里我不知道如何做到这一点，因为我对SQL很陌生。

原始数据

期望输出

任何建议将是非常明显的提前感谢！

Answer 1

您可以使用条件聚合。问题是得到适当的排序和row_number()可以帮助有：

select fieldname, 
     max(case when seqnum = 1 then oldvalue end) as oldvalue, 
     max(case when seqnum = cnt then newvalue end) as newvalue 
from (select t.*, 
      row_number() over (order by id) as seqnum, 
      count(*) over() as cnt 
     from t 
    ) t 
group by fieldname; 

Answer 2

假设你的id是和增量INT（即身份）

Declare @Table table (id int,FieldName varchar(25),OldValue varchar(20),NewValue varchar(25)) 
Insert into @Table (id,FieldName,OldValue,NewValue) values 
(1,'MI','M','N'), 
(2,'GENDER','Male','Female'), 
(3,'GENDER','Female','Male'), 
(4,'FIRSTNAME','','J'), 
(5,'FIRSTNAME','J','Joh'), 
(6,'FIRSTNAME','Joh','John'), 
(7,'STATE','','NJ'), 
(8,'COUNTRY','US','USA') 


Select Sequence = min(MinRow) 
     ,A.FieldName 
     ,OldValue = max(case when id=B.MinRow then OldValue else null end) 
     ,NewValue = max(case when id=B.MaxRow then NewValue else null end) 
From @Table A 
Join (Select FieldName,MinRow=min(ID),MaxRow=max(ID) From @Table Group by FieldName) B on A.FieldName=B.FieldName 
Group by A.FieldName 
Order by 1 

返回

Sequence FieldName OldValue NewValue 
1   MI   M   N 
2   GENDER  Male  Male 
4   FIRSTNAME    John 
7   STATE     NJ 
8   COUNTRY  US   USA 

添加序列并按顺序排列如此

Answer 3

这应该产生正确的结果使用相当b ASIC SQL：

SELECT 
    mc1.FieldName, 
    mc2.OldValue, -- OldestValue would perhaps be more meaningful 
    mc3.NewValue -- Likewise, NewestValue here. 
FROM (
    SELECT FieldName, MIN(Id) MinId, MAX(Id) MaxId 
    FROM ModelChange 
    GROUP BY FieldName 
) mc1 
INNER JOIN ModelChange mc2 ON mc1.MinId = mc2.Id 
INNER JOIN ModelChange mc3 ON mc1.MaxId = mc3.Id; 

我所做的假设，即您的ID是渐进的，对于一个给定的字段名的最早的值将具有最小ID和最新的值将具有最大编号。