1. Take the first N records
Oracle：Select * from TableName where rownum <= N;
DB2：Select * from TableName fetch first N rows only;
2. Get the system date
Oracle：Select sysdate from dual;
DB2：Select current timestamp from sysibm.sysdummy1;
3. Null value conversion
Oracle：Select productid,loginname,nvl(cur_rate,'0') from TableName ;
DB2：Select productid,loginname,value(cur_rate,'0') from TableName;
Coalesce(cur_rate,'0')
4. Type conversion ( version 8 has to_char, to_date, and version 9 adds to_number )
Oracle：select to_char(sysdate,'YYYY-MM-DD HH24:MI:SS') from dual;
DB2：select varchar(current timestamp) from sysibm.sysdummy1;
Oracle data type changing functions: to_char() , to_date() , to_number() , etc.; if you only take the year, month, day, etc., you can use to_char(sysdate, 'YYYY'), to_char('MM') , to_char(' DD') to obtain. Only take the year, month, and day TRUNC(SYSDATE) , and take the hour, minute, and second TO_CHAR(SYSDATE,'HH24:MI:SS') .
DB2 data type changing functions: char() , varchar() , int() , date() , time() , etc.; get the writing of year, month, day, etc.: YEAR(current timestamp) , MONTH(current timestamp) , DAY (current timestamp) , HOUR(current timestamp) , MINUTE(current timestamp) , SECOND(current timestamp) , MICROSECOND(current timestamp) , you can use DATE(current timestamp) only for the year, month and day, and TIME(current timestamp) for the time, minute and second . Char() is a fixed-length string ( 1-255 ), varchar() is a non-fixed-length string ( 1-32672 ) date ,Change the time form to character form : char(current date) , char(current time) convert the string to date or time form : TIMESTAMP('2002-10-20 12:00:00'),DATE('2002-10 -20'),DATE('10/20/2002'),TIME('12:00:00')
Currently DB2 V8 also supports to_char and to_date
5. Quickly clear large tables
Oracle：truncate table TableName ;
DB2：alter table TableName active not logged initially with empty table;
6. About ROWID
Oracle It is uniquely generated by the database and can be obtained in the program
DB2 v8 also has this feature.
7、To_Number
Oracle：select to_number('123') from dual;
DB2： select cast('123' as integer) from sysibm.sysdummy1;
SELECT CAST ( current time as char(8)) FROM sysibm.sysdummy1
8. Create a similar table
Oracle：create table a as select * from b ;
DB2：create table a like b ;
CREATE TABLE tab_new AS select col1,col2 … FROM tab_old DEFINITION ONLY ( valid for version 8, invalid for version 9 )
9. decode method _
Oracle：decode方法（DECODE(条件,值1,翻译值1,值2,翻译值2,...值n,翻译值n,缺省值)）或者case语句
DB2中只有CASE表达式
SELECT id ,name ,
CASE
WHEN integer(flag)=0 THEN ‘假’
WHEN integer(flag)=1 THEN ‘真’
ELSE ‘异常’
END
FROM TEST
或者
SELECT id ,name ,
CASE integer(flag)
WHEN 0 THEN ‘假’
WHEN 1 THEN ‘真’
ELSE ‘异常’
END
FROM TEST
10、子查询（8版，9版也支持子查询）
Oracle: 直接用子查询
Db2: with语句
WITH a1 AS
(select max(id) as aa1 from test )
select id ,aa1 from test ,a1
11、数据类型
比较大的差别：
Oracle：char 2000
DB2: char 254
Oracle: date datetime
Db2: DATE：日期 ，TIME：时间， TIMESTAMP：日期时间
http://hi.baidu.com/neusunly/blog/item/3867a9faaae96fdcb58f3148.html
1、数据类型转换函数
| 整型转字符型 |
字符串转整形 |
字符串转浮点型 |
浮点型转字符串 |
字符串转日期 |
字符串转时间戳 |
日期转字符串 |
|
| ORACLE |
to_char(1) |
to_number('1') |
to_number('1.1') |
to_char(1.1) |
to_date('2007-04-26','yyyy-mm-dd') |
to_date('2007-04-26 08:08:08','YYYY-MM-DD HH24:MI:SS') |
to_char(to_date('2007-04-29','yyyy-mm-dd'),'yyyy-mm-dd') |
| DB2 |
char(1) |
int('1') |
double('1.1') |
char(1.1) |
date('2007-04-26') |
to_date('2007-04-26 08:08:08','YYYY-MM-DD HH24:MI:SS') |
char(date('2007-04-29')) |
| 兼容写法 |
cast(1 as char) |
cast('1' as int) |
无 |
无 |
无 |
兼容 |
无 |
2、Where条件弱类型判断
oracle: where 字符型字段 in (整形) 是允许，DB2不允许
select 'abc' from dual where '1' in (1) 在oracle下可通过
select 'abc' from sysibm.sysdummy1 where '1' in (1) 在DB2下报错
oracle：where 字符型字段=数字型字段 允许，DB2不允许
select 'abc' from dual where '1'=1 在oracle下可通过
select 'abc' from sysibm.sysdummy1 whre '1'=1 在DB2下报错
3、replace关键字
oracle支持，DB2不支持 create or replace语句在DB2下是非法的
4、子查询别名
ORACLE 支持select * from(select 1 from dual) 或者 select * from(select 1 from dual) t
DB2 支持select * from(select 1 from sysibm.sysdummy1) t 或者 select * from(select 1 from sysibm.sysdummy1) as t
固兼容的写法是select * from(子查询) t
5、DATE数据类型的区别
ORACLE中DATE型也是带有时分秒的，但DB2下DATE只是年月日，如'2007-04-28'，且可作为字符串直接操作，DB2中要记录时分秒必须采用TIMESTAMP型
一个采用hibernate后常见的兼容问题是：
如果在映射文件中定义了某个字段为Date型,
则在DB2下，此字段必须定义为timestamp，而不能定义成DATE，不然会报出字符串右截断的错误
对于DB2来说，在查询条件中可以直接用字符串指定日期或时间戳类型字段的值，例如 where create_date = '2007-04-26' 、where create_timestamp = '2007-04-26 08:08:08' ，无须使用字符串转日期函数
6、分页的处理
如果采用JDBC分页的话，注意rownum在DB2中不受支持，比如从masa_area表中取得area_id最小的10条记录，语句分别如下，注意这里的别名t书写方法
ORACLE: select t.* from (select rownum as r1 ,masa_area.* from masa_area order by area_id) t where t.r1<=10
DB2: select t.* from (select rownumber() over() as r1 ,masa_area.* from masa_area order by area_id) t where t.r1<=10
7、decode函数
decode函数在DB2不被支持，兼容的写法是采用case when
8、NVL函数
nvl写法在DB2不被支持，兼容的写法是采用coalesce
ORACLE: select NVL(f_areaid,'空') from masa_user 等同于 select coalesce(f_areaid,'空',f_areaid) from masa_user
DB2: select coalesce(f_areaid,'空',f_areaid) from masa_user
9、substr的不同
DB2 substr举例如下：
masa_group表的f_groupCode字段定义成VARCHAR(100),所以下面这个语句不会出错，如果是substr(f_groupCode,1,101)就出错了
select * from masa_group where substr(f_groupCode,1,50) = '001006' order by f_groupcode
在DB2下无错，但是
select * from masa_group where substr('001006', 1, 50) = '001006' order by f_groupcode
就报错，说第三个参数超限
这是因为'001006'已经定义为一个长度为6的charater了
这点和ORACLE有很大不同，请大家注意
如果是要从第一位取到最后一位，稳妥的办法是不要加第三个参数
ORACLE：select substr('123456',1) from dual
DB2：select substr('123456',1) from sysibm.sysdummy1
都没有问题
from:
http://wjxk.blog.sohu.com/82119411.html
比较全：
http://blog.itpub.net/14075938/viewspace-475099/