临时表中行列直接的装换

Posted 2023-05-08

该问题的关键是SQL Server 2000字段大小限制了。
有个笨的办法，将325个物料分成两组（如果还不够，分成3组），用你这种方法分别进行行列转换，放入临时表，然后将临时表连接起来。

如果是SQL Server 2005就好办了，直接将@sql申明成varchar(max)就行了，它最大支持2GB,应该可以解决此问题。另外，SQL Server 2005专门针对行列转换，有了pivot子句。
参考技术A pivate 与 unpivate 这两个关键字，就是用来将行转换成列，把列转换成行的。

行列转换

 1 /****** Object:  Table [dbo].[Table_1]    Script Date: 08/06/2013 13:55:39 ******/  
 2 /*行列转换*/
 3 
 4 SET ANSI_NULLS ON  
 5 GO  
 6   
 7 SET QUOTED_IDENTIFIER ON  
 8 GO  
 9   
10 SET ANSI_PADDING ON  
11 GO  
12 
13 /*对临时表无效*/  
14 IF OBJECT_ID (‘#Table_1‘, ‘U‘) IS NOT NULL  
15    DROP TABLE #Table_1;  
16 
17 CREATE TABLE #Table_1(  
18     [name] [varchar](50) NOT NULL,  
19     [score] [real] NOT NULL,  
20     [subject_id] [int] NOT NULL  
21 ) ON [PRIMARY]  
22   
23 insert into #Table_1 ([name],[score],[subject_id]) values(  ‘a‘    ,   90  ,   1   );  
24 insert into #Table_1 ([name],[score],[subject_id]) values(  ‘b‘    ,   80  ,   2   );  
25 insert into #Table_1 ([name],[score],[subject_id]) values(  ‘c‘    ,   70  ,   3   );  
26 insert into #Table_1 ([name],[score],[subject_id]) values(  ‘d‘    ,   50  ,   1   );  
27 insert into #Table_1 ([name],[score],[subject_id]) values(  ‘e‘    ,   40  ,   2   );  
28 insert into #Table_1 ([name],[score],[subject_id]) values(  ‘f‘    ,   60  ,   1   );  
29 
30 --显示1
31 
32 SELECT [name],[1],[2],[3]  
33 FROM #Table_1  
34 pivot  
35 (  
36   sum(score) for subject_id in ([1],[2],[3])  
37 ) as pvt  
38 
39 select * from #Table_1
40 
41 --显示2
42 
43 SELECT [name],[subject_id],[score]  
44 FROM  
45 (  
46     SELECT [name],[1],[2],[3]  
47     FROM #Table_1
48     pivot  
49     (  
50       sum(score) for subject_id in ([1],[2],[3])  
51     ) as pvt  
52 ) p  
53 unpivot  
54 (  
55     score for subject_id in([1],[2],[3])  
56 ) as unpvt