TSQL update 简单应用小总结

UPDATE 有两种基本的格式。一种是用静态数据来修改表，另一种是用其他表中的数据来修改表。下面是第一种格式：

1 UPDATE #famousjaycees

2    SET jc = 'Jhony cash',

3        occupation = 'singer/songwrite',

4        becamefamous = 1955,

5        notes = 'began c'

6  WHERE jc = 'Jhony caeer'

第二种格式如下：

 1 IF NOT EXISTS

 2       (SELECT *

 3          FROM sys.sysobjects

 4         WHERE [name] = 'semifamousjaycees')

 5    CREATE TABLE semifamousjaycees

 6    (

 7       jc             VARCHAR (15),

 8       occupation     VARCHAR (25),

 9       becamefamous   INT DEFAULT 0,

10       notes          TEXT NULL

11    )

12 

13 UPDATE f

14    SET jc = s.jc,

15        occupations = s.occupations,

16        becamefamous = s.becamefamous,

17        notes = s.notes

18   FROM famousejaycees f

19        JOIN semifamousjaycee ON (f.becamesou = s.becamefouse)

UPDATE 的另外一种用途：使用update 检测约束。

如果使用 bulk insert 或者其他大批量的载入工具来对有insett 触发器的表进行追加数据，你们你会发现触发器不能触发。而且，即使bulk insert 不妨碍约束，也会是操作变的非常的慢。如果在载入数据时忽略约束，那么就快的多了。所以，比较的好的方法是，载入数据结束后，马上在对表一个假的update 操作。这个假的修改操作只是简单地将列值为其身的值。这样就会触发触发器约束进行检测。如果其中有包含错数数据的行，那么update失败.例如：

 1 IF NOT EXISTS

 2       (SELECT *

 3          FROM sys.sysobjects

 4         WHERE [name] = 'famousejaycees')

 5    CREATE TABLE famousejaycees

 6    (

 7       jc             VARCHAR (15) CHECK (left (jc, 3) <> 'Joe'), -- establish a check constraint

 8       occupation     VARCHAR (25),

 9       becamefamous   INT DEFAULT 0,

10       notes          TEXT NULL

11    )

12 go

13 

14 BULK INSERT famousejaycees

15        FROM 'C:\famous.bcp'

16        WITH (FIELDTERMINATOR = '|', ROWTERMINATOR = '\r\n')

17 

18 --check that the miscreant is in place

19 SELECT * FROM famousejaycees

20 

21 --now do the faux update

22 

23 update famousejaycees

24 set jc=jc, occupation = occupation, becamefamous = becamefamous, notes = notes

使用update 交换行列：

IF NOT EXISTS

      (SELECT *

         FROM sys.sysobjects

        WHERE [name] = 'sample')

   CREATE TABLE sample

   (

      k1      INT IDENTITY,

      samp1   FLOAT DEFAULT (rand () * 1000),

      samp2   FLOAT DEFAULT (rand () * 1000)

   )



SELECT * FROM sample;



DECLARE @swap   FLOAT



UPDATE sample

   SET @swap = samp1, samp1 = samp2, samp2 = @swap