原帖:http://community.csdn.net/Expert/topic/5403/5403570.xml?temp=.6892511
在回帖中发现有一方法不错,那就是 marco08(天道酬勤) 的回帖,如下:
1
create
table
T(A
decimal
(
10
,
1
), B
decimal
(
10
,
1
), C
decimal
(
10
,
1
), D
decimal
(
10
,
1
), E
decimal
(
10
,
1
))
2 insert T select - 21.5 , - 15.0 , - 5.0 , null , null
3 union all select - 5.5 , - 11.5 , null , null , null
4 union all select - 1.0 , - 16.5 , - 10.5 , null , null
5
6
7 select * ,
8 max_value = (
9 select max (A) from
10 (
11 select A
12 union all
13 select B
14 union all
15 select C
16 union all
17 select D
18 union all
19 select E
20 )tmp)
21 from T
22
2 insert T select - 21.5 , - 15.0 , - 5.0 , null , null
3 union all select - 5.5 , - 11.5 , null , null , null
4 union all select - 1.0 , - 16.5 , - 10.5 , null , null
5
6
7 select * ,
8 max_value = (
9 select max (A) from
10 (
11 select A
12 union all
13 select B
14 union all
15 select C
16 union all
17 select D
18 union all
19 select E
20 )tmp)
21 from T
22
--result
A B C D E max_value
------------ ------------ ------------ ------------ ------------ ------------
-21.5 -15.0 -5.0 NULL NULL -5.0
-5.5 -11.5 NULL NULL NULL -5.5
-1.0 -16.5 -10.5 NULL NULL -1.0
(3 row(s) affected)
这一方法,自我感觉不错,还真的第1次看到这样的写法。原来SQL里面还可以实现这样的写法,又学到了一点知识。