亚洲AV永久无码天堂影院,亚洲va久久久噜噜噜久久男同 ,国产亚洲综合色就色

Jcat — Fri, 17 Jun 2011 07:35:00 GMT

��单过一遍，详细的以后遇到慢慢研�I��?br />主要的更新在几个斚w��Q?br />1. OUI的改�q?br />2. Oracle Restart
3. ASM的增强！�Q�！

New Grid Infrastructure Installation Option
1. 单点包括�Q�ASM、listener和Oracle Restart�Q�监控、管理�ƈ自动重启各个�l��g�Q?br /> 2. 集群包括�Q�ASM、listener和Clusterware

New Desktop and Server Class Options
1. Desktop Class�Q�适用于笔记本、台式机
2. Server Class�Q�适用于服务器�Q�功能上没区别，多一些高�U�配�|�选项�Q?br />
Daylight Savings Time Upgrade of Timestamp with Timezone Data Type
新的DBMS_DST包，优化对TIMESTAMP WITH TIMEZONE数据的管�?br />
SYSASM Privilege
��理ASM需要SYSASM权限�Q�旨在分清ASM��理和DB��理

Fixup Scripts and Prerequisite Checks
安装前检查时�Q�如果遇��C��W�合要求的配�|�，OUI会对一些检查项自动生成fixup脚本�Q�用root执行��可以解决相应问题�?br />
New Tool to Configure Custom Installation Options
OUI不在提供对单个组件的配置功能�Q�如果需要只能用$ORACLE_HOME/bin/chopt命��o行进行配�|��?b>我想Oracle的趋势就是�ؓ了让更多已经成熟�?font color="#ff0000">配置自动�?/font>�Q�简化��用�?/b>

Deinstallation Tool
OUI不再用来删除oracle软�g�Q�请使用$ORACLE_HOME/deinstall

Intelligent Data Placement
指定ASM��盘的磁盘范��_��旨在��高频访问的数据攑֜�HOT region�Q�比如磁盘的外道�Q?br />
Oracle Automatic Storage Management Cluster File System (Oracle ACFS)
为其它文件提供ASM的存储功�?br />
Data Pump Export and Data Pump Import
Data Pump兼容Export/Import

Use Oracle Restart to Automatically Restart Your Database
1. 针对单点环境
2. 自动重启又问题的�l��g�Q�如�Q�db instance、listener、asm instance

New Method of Installing Oracle Automatic Storage Management
以前�Q�ASM的安装是伴随着DB的安装；现在是伴随着Software的安装�?br /> 其实�Q�就是把ASM上升��C��个��Y件的高度�Q�以前只是db的一个组�Ӟ��

SRVCTL Support for Single Instance Database in a Cluster
SRVCTL�l�一��理单点�Q�with Restart�Q�和集群数据库（with Clusterware�Q?br />
~~Deprecated in Oracle Database 11g Release 2~~
不支持裸讑֤�了：要么用文件系�l�、要么用ASM�Q�说白了�Q?font color="#ff0000">企业环境��必��ȝ��ASM
不支持Oracle Ultra Search�Q?i的一个什么鸟功能�Q?br />

Jcat 2011-06-17 15:35 发表评论

Win - 启动�l�色版MYSQL

Jcat — Wed, 15 Jun 2011 17:40:00 GMT

重装完系�l�，mysql不需要重装，直接当绿色版使用�Q�还更干净

//启动MYSQL
cd $MYSQL_HOME/bin
D:\JAVA\MYSQL\mysql-5.1.53-win32\bin>mysqld --console
110616 1:26:26 [Note] Plugin 'FEDERATED' is disabled.
InnoDB: The log sequence number in ibdata files does not match
InnoDB: the log sequence number in the ib_logfiles!
110616 1:26:26 InnoDB: Database was not shut down normally!
InnoDB: Starting crash recovery.
InnoDB: Reading tablespace information from the .ibd files...
InnoDB: Restoring possible half-written data pages from the doublewrite
InnoDB: buffer...
110616 1:26:26 InnoDB: Started; log sequence number 0 44233
110616 1:26:27 [Note] Event Scheduler: Loaded 0 events
110616 1:26:27 [Note] mysqld: ready for connections.
Version: '5.1.53-community' socket: '' port: 3306 MySQL Community Server (GPL
)

//��单��?/font>
mysql> show databases;
+--------------------+
| Database           |
+--------------------+
| information_schema |
| mysql              |
| test               |
+--------------------+
3 rows in set (0.03 sec)

mysql> use mysql
Database changed

mysql> show tables;
+---------------------------+
| Tables_in_mysql           |
+---------------------------+
| columns_priv              |
| db                        |
| event                     |
| func                      |
| general_log               |
| help_category             |
| help_keyword              |
| help_relation             |
| help_topic                |
| host                      |
| ndb_binlog_index          |

登陆
D:\JAVA\MYSQL\mysql-5.1.53-win32\bin>mysql.exe -uroot
创徏数据�?/b>
mysql> create database testdb;

Jcat 2011-06-16 01:40 发表评论

Jcat — Wed, 15 Sep 2010 07:26:00 GMT
--主键是非分区索引�Q�也可以看作是全局
create table test_par1
(
tdate   varchar2(8) primary key
)
partition by range ( tdate )
(
     partition p1 values less than ('20090201'),
     partition p2 values less than ('20090301'),
     partition pm values less than (MAXVALUE)
) tablespace test;

--主键是分区烦�?/font>
create table test_par2
(
tdate   varchar2(8)
)
partition by range ( tdate )
(
     partition p1 values less than ('20090201'),
     partition p2 values less than ('20090301'),
     partition pm values less than (MAXVALUE)
) tablespace test;

create index i_tdate2 on test_par2(tdate) local;

alter table test_par2 add constraint pk_tdate2 primary key(tdate);

--查看test_par2的DDL
select dbms_metadata.get_ddl( 'TABLE', 'TEST_PAR2' ) from dual;
--整理如下
create table test_par3
(
tdate   varchar2(8),
constraint pk_tdate3 primary key (tdate) using index local
)
partition by range ( tdate )
(
     partition p1 values less than ('20090201'),
     partition p2 values less than ('20090301'),
     partition pm values less than (MAXVALUE)
) tablespace test;

�q�个案例也告诉我们，在做DDL�Ӟ��量�q�是昄��的写出属性，一些简易语法会引�v不可知的定义。最后都�?get_ddl 再查一下，�q�才是Oracle真正执行的DDL�?/font>

Jcat 2010-09-15 15:26 发表评论

DB2基础命��o

Jcat — Fri, 09 Jul 2010 06:32:00 GMT
列出所有实�?/font>
[db2inst1@dcm ~]$ db2ilist
db2inst1

列出当前实例下的所有数据库
[db2inst1@dcm ~]$ db2 list db directory

System Database Directory

Number of entries in the directory = 1

Database 1 entry:

Database alias                       = SAMPLE
Database name                        = SAMPLE
Local database directory             = /home/db2inst1
Database release level               = d.00
Comment                              =
Directory entry type                 = Indirect
Catalog database partition number    = 0
Alternate server hostname            =
Alternate server port number         =

启动实例
[db2inst1@dcm ~]$ db2start
07/09/2010 14:24:36     0   0   SQL1063N DB2START processing was successful.
SQL1063N DB2START processing was successful.

�q�接到数据库
[db2inst1@dcm ~]$ db2 connect to sample

   Database Connection Information

Database server        = DB2/LINUX 9.7.1
SQL authorization ID   = DB2INST1
Local database alias   = SAMPLE

执行SQL
[db2inst1@dcm ~]$ db2
db2 => select * from staff where dept=20

ID     NAME      DEPT   JOB   YEARS SALARY    COMM
------ --------- ------ ----- ------ --------- ---------
    10 Sanders       20 Mgr        7 98357.50         -
    20 Pernal        20 Sales      8 78171.25    612.45
    80 James         20 Clerk      - 43504.60    128.20
   190 Sneider       20 Clerk      8 34252.75    126.50

4 record(s) selected.

退�?/font>
db2 => quit
DB20000I The QUIT command completed successfully.

断开�q�接�Q�不断开��׃��能db2stop�Q?/font>
[db2inst1@dcm ~]$ db2 connect reset
DB20000I The SQL command completed successfully.

关闭实例
[db2inst1@dcm ~]$ db2stop
07/09/2010 14:29:32     0   0   SQL1064N DB2STOP processing was successful.
SQL1064N DB2STOP processing was successful.

Jcat 2010-07-09 14:32 发表评论

数据库、实�?on Oracle/DB2

Jcat — Thu, 08 Jul 2010 10:26:00 GMT
最�q�初探了一下DB2�Q�发现关于“数据库”、“实例”的概念及其关系�Ҏ��把�h弄晕�Q�这里小��ȝ��一下，�q�把Oracle也拉�q�来一��h��?br />另外�Q�发现Oracle和IBM老搞一些理��Z��针锋相对的事情�Q�如RAC vs DPF (database partitioning feature)�Q�很是有��?/font>

基本概念�q�是相同�?/b>
数据库��Y�Ӟ��下面��U?b>软�g�Q?/font>�Q��Y件本�w�，一��C��机安装一套就可以了。（抛开你故意在一台服务器上安�?遍Oracle/DB2�q�种不正常思维�Q?br />
实例�Q�数据库软�g启动后的“进�E?内存�?br />
软�g+实例的关�p�：一台电脑只需要按照一套QQ�Q��Y�Ӟ��Q�但是我们可以多用户��d��Q�实例）。换到哲学的角度�Q�就是抽象和具体的关�p�R�?br />
数据库（狭义�Q�下面简�U?b>�?/b>�Q?/font>�Q�一堆文�Ӟ��数据文�g、控制文件、日志文件、参数文�Ӟ��
它和软�g的区别：软�g是厂商卖�l�你的东西；�?/b>是你自己的东西（业务、应用）
它和实例的关�p�：用户需要借助实例�Q�所提供的各�U�手�D�）来访�?b>�?/b>
�?/b>��q��当于是QQ聊天记录

最后：数据库（�q�义�Q? 软�g + 实例 + �?/font>
一台服务器�Q�即一套数据库软�g�Q�可以徏多个实例�Q�多个库�Q�且互不相干

体系�l�构的不同之�?/b>
Oracle 实例和库一一对应
DB2     一个实例可以挂多个�?br />

高��情况�Q�简单提一下，以后再慢慢研�IӞ��
Oracle�Q�RAC一个库被多个实例挂
DB2�Q�DPF一个实例多个库
在数据库集群模式设计斚w��Q�有Share everything架构和Share nothing架构�Q�前者以ORACLE RAC��Z��表，IBM DB2 Purescale为挑战者；后者以IBM DB2 DPF��Z��表�?br />

Jcat 2010-07-08 18:26 发表评论

Jcat — Sun, 13 Jun 2010 06:08:00 GMT
行链接：
1. 一条记录的大小大于block size�Q�则产生行链�?br />2. �Ҏ��发生在比较大的行�?br />3. 因�ؓ行链接是由db_block_size不够大引��L��Q�所以对已有的行链接是无法清除的
4. 9i以后�Q�可以对不同的表�I�间讄��不同的db_block_size�Q�可以将一些特�D�的宽表攑֜�大block size的表�I�间

例子�Q?br />表空间block size�?k�Q?192�Q�，因�ؓ数据块头也要占一定空��_��所以如下例�Q�实际只能放7948的数据，一旦超�q�，��׃�生行链接

--�?br />create table test7948(a char(2000),b char(2000),c char(2000),d char(1948))
tablespace test;

insert into test7948 values('a','b','c','d');
commit;

--�?br />create table test7949(a char(2000),b char(2000),c char(2000),d char(1949))
tablespace test;

insert into test7949 values('a','b','c','d');
commit;

行迁�U�：
1. 本来是放的下�?br />2. 因�ؓ更新使row size变大了，一个block里又不��以放下增加的�I�间�Q�PCTFREE相关�Q�，则��生行�q�移
3. �Ҏ��发生在PCTFREE较小�Q�对�c�M��varchar�c�d��的update又很多的表上
4. 无法避免�Q�但通过把数据导出导入进行清�?br />
例子�Q?br />--�?br />create table test7948_vchar(a char(2000),b char(2000),c char(2000),d char(1940), e varchar(9))
tablespace test;

insert into test7948_vchar values('a','b','c','d','12345678');
commit;

一更新�Q��得row size大于7948了，产生行迁�U?br />--�?br />update test7948_vchar set e='123456789'

Jcat 2010-06-13 14:08 发表评论

Jcat — Fri, 30 Apr 2010 03:05:00 GMT
--创徏一个大一点的TEMP表空�?br />create temporary tablespace temp1
tempfile '/oracledatafile/temp01.dbf' SIZE 100m autoextend on next 100m maxsize 5000m;

--切换默认TEMP表空�?br />alter database default temporary tablespace temp1;

--删掉以前的TEMP表空�?br />drop tablespace temp including contents and datafiles;

Jcat 2010-04-30 11:05 发表评论

物化视图

Jcat — Tue, 12 Jan 2010 16:04:00 GMT
1. 物化视图�׃��是物理真实存在的�Q�故可以创徏索引�?

--为基表创建MLOG
--创徏物化视图时应先创建存储的日志�I�间�Q�否则徏MV时报�?br />--ORA-23413: table "SCOTT"."EMP" does not have a materialized view log
create materialized view log on scott.emp
tablespace test
/

--创徏物化视图
create materialized view test_mv
tablespace test
parallel (degree 4)
build immediate refresh fast
enable query rewrite
as
select * from scott.emp
/

--查看一下结果，果然很符合物化视囄��定义�Q�一个表+一个视�?/font>
SQL> select object_name, object_type from user_objects where object_name = 'TEST_MV';
OBJECT_NAME    OBJECT_TYPE
-----------    --------
TEST_MV        TABLE
TEST_MV        MATERIALIZED VIEW

SQL> select mview_name, container_name from user_mviews;
MVIEW_NAME       CONTAINER_NAME
---------------- ------------------------------
TEST_MV          TEST_MV �Q�这��是那个存储表）

--查看MLOG的情�?/font>
--注意�Q�MLOG的所属和MV的所属�ƈ不是同一�?/font>
SQL> select log_owner, master, log_table from dba_mview_logs
LOG_OWNER    MASTER    LOG_TABLE
---------------------------------
SCOTT        EMP    MLOG$_EMP �Q�MLOG其实也就是一个表�Q?/font>

SQL> desc scott.mlog$_emp;
Name            Type
-------------------------------
EMPNO            NUMBER(4)
SNAPTIME$$        DATE
DMLTYPE$$        VARCHAR2(1)
OLD_NEW$$        VARCHAR2(1)
CHANGE_VECTOR$$        RAW(255)

--删除MLOG
drop materialized view log on 物化视图所依赖的表�?
--删除物化视图
drop materialized view 物化视图�?

Jcat 2010-01-13 00:04 发表评论

Jcat — Mon, 11 Jan 2010 06:58:00 GMT
注意�Q��ؓ了能在终端看见put_line的输出，�q�需要先开�?br />set serveroutput on

--用来��试的表
create table test_tri(
a1 int,
a2 int
)
tablespace test

-----行��insert触发�?/font>-----
1) 插入的一行新数据保存�?b>:new
2) insert触发器没�?old�?/font>
3) �?new的修改，只能定义在before�c�d��的触发器�?/font>

---触发器语�?--
create or replace trigger 名称
[after|before] [delete|update|insert]
[of 列名] [on 表名]
[referencing new as 新行别名 old as 旧行别名]
[for each row] [when(条�g)]
declare
....
begin
...
exception
....
end;

--定义
create or replace trigger test_before_insert
before insert on test_tri
for each row when(new.a2 is null)
begin
   dbms_output.put_line('insert row original: a1=' || :new.a1 || ' a2=' || :new.a2);
   :new.a2 := :new.a1 * 2;
end;
/

create or replace trigger test_after_insert
after insert on test_tri
for each row
begin
   dbms_output.put_line('insert row actual: a1=' || :new.a1 || ' a2=' || :new.a2);
end;
/

--��试
SQL> insert into test_tri(a1) values(1);
insert row original: a1=1 a2=
insert row actual: a1=1 a2=2

SQL> insert into test_tri values(2,3);
insert row actual: a1=2 a2=3

SQL> select * from test_tri;
        A1         A2
---------- ----------
         1          2(��p��发器生成的�?
         2          3

-----DDL触发�?----
--��M��create语句都会触发�q�个语句
create or replace trigger test_ddl_trigger
before create on schema
begin
    dbms_output.put_line( 'DDL Trigger' );
    insert into test_tri values(9,9);
end;

Jcat 2010-01-11 14:58 发表评论

IPC extproc PLSExtPro

Jcat — Tue, 05 Jan 2010 13:59:00 GMT
Oracle建好后，tnsnames和listener中自动就带有如下内容�Q�这里咱们就来解释一下这些东西是�q�什么用�?/font>

----TNSNAMES.ora----
EXTPROC_CONNECTION_DATA =
(DESCRIPTION =
    (ADDRESS_LIST =
      (ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC1))
    )
    (CONNECT_DATA =
      (SID = PLSExtProc)
      (PRESENTATION = RO)
    )
)

----LISTENER.ora----
SID_LIST_LISTENER =
(SID_LIST =
    (SID_DESC =
      (SID_NAME = PLSExtProc)
      (ORACLE_HOME = /opt/oracle/10gR2)
      (PROGRAM = extproc)
    )
)

LISTENER =
(DESCRIPTION_LIST =
    (DESCRIPTION =
      (ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC1))
      (ADDRESS = (PROTOCOL = TCP)(HOST = dcm)(PORT = 1521))
    )
)

IPC - Inner Process Communication
When a process is on the same machine as the server, use the IPC protocol for connectivity instead of TCP. Inner Process Communication on the same machine does not have the overhead of packet building and deciphering that TCP has.
I've seen a SQL job that runs in 10 minutes using TCP on a local machine run as fast as one minute using an IPC connection. The difference in time is most dramatic when the Oracle process has to send and/or receive large amounts of data to and from the database.
For example, a SQL*Plus connection that counts the number of rows of some tables will run about the same amount of time, whether the database connection is made via IPC or TCP. But if the SQL*Plus connection spools much data to a file, the IPC connection will often be much faster -- depending on the data transmitted and the machine workload on the TCP stack.

For how to configure it:
1. you should add one IPC line in the LISTENER.ORA
2. You should also add one IPC line in the TNSNAMES.ORA

PLSExtPro - PL/Sql External Procdure
默认安装�Ӟ��会安装一个PL/SQL外部�E�序�Q?font color="#ffa500">extproc--�q�是�E�序名）条目在listener.ora中，是oracle��用外部程序默认配�|�的监听�Q�它的名字（也就是SID�Q�通常是ExtProc�?font color="#ff0000">PLSExtProc�?br />但一般不会��用它�Q�可以直接从listener.ora中将�q�项�U�除�Q�因为对ExtProc已经有多�U�攻��L��D�了�Q�在不��用外部程序时�Q�Oracle也是��删除的�?br />

extproc的作�?/font>��是在pl/sql中调�?b>外部语句�Q�如c�Q�java写的�q�程�?br />现在�Q�Oracle已经全面支持JAVA了，�q�东西也��p��时了�Q�之所以��l�保留是考虑到兼容以前老版本的数据库实例�?br />
[oracle@dcm bin]$ extproc
Oracle Corporation --- TUESDAY   JAN 05 2010 21:58:23.878
Heterogeneous Agent Release 10.2.0.1.0 - Production

Jcat 2010-01-05 21:59 发表评论

Oracle Scheduler 10G NEW

Jcat — Thu, 17 Dec 2009 08:13:00 GMT
--��d��sys用户�Q�创��Z��个测试用��P��权限可以�l�大�?/font>
SQL> create user test identified by xxxxx;
SQL> grant connect to test;
SQL> grant resource to test;
SQL> grant dba to test;

--��d��test用户�Q�开始测�?/font>
--��Z��试�?/font>
create table test_table(id int, time timestamp);

--创徏Job
begin
dbms_scheduler.create_job(
   job_name => 'test_job',
   job_type => 'PLSQL_BLOCK',
   job_action => 'insert into test_table
                  (select nvl(max(id),0)+1, systimestamp from test_table, dual);', --nvl函数同SQLServer的isnull函数
   start_date => null, --一�Ȁ�z�，��开�?/font>
   repeat_interval => 'FREQ=SECONDLY;INTERVAL=10');
end;

FREQ用来指定间隔的时间周期，可选参数有�Q�YEARLY�Q�MONTHLY�Q�WEEKLY�Q�DAILY�Q�HOURLY�Q�MINUTELY�Q�SECONDLY�?/font>

--光创建没用，�q�需要激�z�；也可以在创徏�Ӟ��直接把enable属性设�|��ؓtrue�Q?font color="#a52a2a">enabled => true�Q?/font>
select job_name, enabled, run_count from user_scheduler_jobs;
JOB_NAME                       ENABL RUN_COUNT
------------------------------ ----- ----------
TEST_JOB                       FALSE          0

begin
dbms_scheduler.enable('test_job');
end;

--查看效果
select id, to_char(time,'HH24:MI:SS:FF3') from test_table;
       ID TO_CHAR(TIME,'HH24
---------- ------------------
         1 16:13:29:542
         2 16:13:39:506
         3 16:13:49:109
         4 16:13:59:097
         5 16:14:09:109
         6 16:14:19:103
         7 16:14:29:101
         8 16:14:39:099
         9 16:14:49:105
        10 16:14:59:100

--停止��d��
begin
dbms_scheduler.disable('test_job');
end;

--删除��d��
begin
dbms_scheduler.drop_job('test_job');
end;

Jcat 2009-12-17 16:13 发表评论

数据库的隔离�U�别

Jcat — Sat, 05 Dec 2009 09:45:00 GMT

一致和�q�发是对立的�Q�需要根据应用，选择权宜之计

数据不一致的现象
---事务内单SQL的情�?--
1.脏读-Dirty Read�Q�本事务��d��了其它事务尚未提交的修改数据
例子�Q�读了不该读�?br />1:00 x=1
1:01 A用户 Update x=2�Q�但未commit�Q?br />1:02 B用户 Select x --> x=2
合理的情冉|��x仍然�{�于1

---事务内多SQL的情况（典型的如�Q�先查再改）---
2.不可重复�?/b>-Non Repeatable Read
例子1�Q�自相矛�?br />1:00 x=1 y=2
1:01 B用户 Select x,y --> x=1 y=2
1:02 A用户 Update x=2; Commit;
1:03 B用户 Select x+y --> x+y=4
首先�q�个�l�果从单条SQL的角度看�Q�是没有问题的。但是，如果把B的两�ơ查询看作一个整体（事务�Q�，那么合理的情况应该是
x+y仍然�{�于3
或者B再进行一�ơ事务，得出 x=2 y=2 x+y=4 的结�?/font>

例子2�Q�更��C��?br />1:00 x=1
1:01 B用户 Select x --> x=1
1:02 A用户 Select x --> x=1
1:03 A用户 Update x=2; Commit;
1:04 B用户 Update x=3; Commit;
同样�Q�从单条SQL来讲�Q�没有�Q何问题�?br />但是从逻辑的合理性讲�Q�一般的更新操作都是先查再改�Q�换�a��?br /> A真正惛_��的是Update x from 1 to 2
B真正惛_��的是Update x from 1 to 3
但最�l�却造成了在B不知情的情况下，把B的初��h��Z��Update x from 2 to 3

3.�q�d��?/b>-Phantom Read
例子�Q�读��C��未来
1:00 X1=1 X2=2
1:01 B用户 Select Xi --> X1=1 X2=2
1:02 A用户 Insert X3=3; Commit;
1:03 B用户 Select sum(Xi) --> re=6
其实道理和之前的不可重复�ȝ��同，只不�q�是由Insert引�v的�Ş了�?br />�Q�甚至Delete也会引�v�c�M��的问题，但好像学术界�q�没有对Delete�q�行讨论�Q?/font>

Isolation Level
Read Uncommitted�Q?�Q?�Q?都会发生
Oracle中严格禁止脏�?br /> 在SQL Server 7.0中，是可以选择该��别的
Read Committed�Q�发�?�Q?�Q�Oracle的默认��别）
Repeatable Read�Q�发�?
Serializable�Q�都不发�?br />

Oracle的实现方�?/b>
Read Committed�Q�默认就实现
Repeatable Read�Q?br /> 1. 悲观锁（select ... for update)�Q�媄响�ƈ�?br /> 2. 乐观锁（update where 所有字�D�都作�ؓ条�g)�Q�不影响�q�发
Serializable�Q?br /> alter session set isolation_level=serializable/read only

Jcat 2009-12-05 17:45 发表评论

Jcat — Tue, 18 Aug 2009 10:17:00 GMT
--创徏一个分��
create table scott.testpart
(
   TEXTDATE   VARCHAR2(8)
)
partition by range ( TEXTDATE )
(
   partition p1 values less than ('20090201'),
    partition "p2" values less than ('20090301'), --注意�q�里的双引号
   partition pm values less than (MAXVALUE)
)

--插入一些数据（��Z��实践一下双层��@环）
DECLARE
   i int:=1;
   j int:=1;
BEGIN
    WHILE i<=3 LOOP
            WHILE j<=3 LOOP
                insert into scott.testpart values (2009*10000+i*100+j);
                j:=j+1;
            END LOOP;
        i:=i+1;
        j:=1; --注意要重�|�j�Q�否则就回不到j循环里面�?/font>
        END LOOP;
END;

--需要先�q�行分析�Q�否则num_rows列将为空
analyze table scott.testpart compute statistics;

--查看各个分区的情�?/font>
select table_name, partition_name, num_rows from DBA_TAB_PARTITIONS
where table_owner='SCOTT' and table_name='TESTPART';
TABLE_NAME                PARTITION_NAME           NUM_ROWS
-------------------------- ------------------------------ ----------
TESTPART                       P1                                            3
TESTPART                       PM                                        3
TESTPART                       p2                                       3
注意p2因�ؓ加了引号仍然是小写（但�ƈ不显�C�引��P��q�点比较讨厌�Q�，没加引号的会昄��为大�?/font>

--验证数据��实�q�入了正��的分区
select * from scott.testpart partition(p1);
TEXTDATE
--------
20090101
20090102
20090103

--创徏p2时有引号�Q��用p2时也��需要引受��?/font>
--特别的，用一些第三方客户端创建分��Ӟ��生成的SQL往往会自动加上引��P��而查看DBA_TAB_PARTITIONS时又看不出来�Q�所以遇��C��面的问题�Q�可以加个引可��试�?/font>
select * from scott.testpart partition(p2);
ERROR at line 1:
ORA-02149: Specified partition does not exist

select * from scott.testpart partition("p2");
TEXTDATE
--------
20090201
20090202
20090203

--其它相关
select * from scott.testpart;
truncate table scott.testpart;
drop table scott.testpart;

Jcat 2009-08-18 18:17 发表评论

NLS_DATE_FORMAT

Jcat — Sat, 27 Jun 2009 08:18:00 GMT
--修改后重新启动数据库�Q�永久生�?/b>
SQL> alter system set nls_date_format='yyyymmdd';
ERROR at line 1: --不能是memory scope�Q�默认是both�Q�所以也不行�Q?/font>
ORA-02096: specified initialization parameter is not modifiable with this option

SQL> alter system set nls_date_format='yyyymmdd' scope=spfile;
System altered.

SQL> show parameter nls_date_format --�q�个时候还没生�?/font>
NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
nls_date_format                      string

--重启数据库后
SQL> show parameter nls_date_format
NAME                                 TYPE        VALUE
------------------------------------ ----------- ------------------------------
nls_date_format                      string      yyyymmdd

SQL> select sysdate from dual;
SYSDATE
--------
20090627

--只对当前session有效�Q�重��d��一�ơSQLPLUS�Q�效果就消失�?/b>
SQL> alter session set nls_date_format='yyyy-mm-dd';   --�q�个格式也很常用 yyyy-mm-dd hh:mi:ss
Session altered. --同时会覆盖初始化参数里的讄��

SQL> select sysdate from dual;
SYSDATE
----------
2009-06-27

Jcat 2009-06-27 16:18 发表评论

Oracle Data Dictionary

Jcat — Thu, 18 Jun 2009 14:42:00 GMT
Oracle所有对象的相关信息都可以通过静态数据字�?/b>来查找，但数据字典实在太多，也记不清名字�?br />因�ؓ数据字典都是�?font color="#ff0000">DBA_开头的视图�Q�所以可以想办法先把它们列出来�?br />btw�Q�DBA>ALL>User

以下两句�l�果上是�{�效�?br />select object_name from dba_objects where object_name like 'DBA\_%' escape '\' and object_type='VIEW';
select view_name from dba_views where view_name like 'DBA\_%' escape '\';

例子�Q?br />我想查看数据文�g的相关信息，但是从DBA_DATA_FILES里，�q�没有看见Temp表空间的数据文�g的信息�?br />于是可以��试如下搜烦�Q?br />SQL> select view_name from dba_views where view_name like 'DBA\_%TEMP%' escape '\';
VIEW_NAME
------------------------------
DBA_ADVISOR_SQLW_TEMPLATES
DBA_ADVISOR_TEMPLATES
DBA_HIST_BASELINE_TEMPLATE
DBA_LOB_TEMPLATES
DBA_REPCAT_REFRESH_TEMPLATES
DBA_SUBPARTITION_TEMPLATES
DBA_TEMPLATE_REFGROUPS
DBA_TEMPLATE_TARGETS
DBA_TEMP_FILES    --扑ֈ�嫌疑犯，�q�去一看，果然记录的是关于Temp表空间的数据文�g的信�?/font>
DBA_TEMP_FREE_SPACE

列一些常用的在这�?/span>

Jcat 2009-06-18 22:42 发表评论

ALTER DATABASE vs ALTER SYSTEM

Jcat — Tue, 16 Jun 2009 04:32:00 GMT
�q�两个命令都是用来更改一些数据库配置的，所以经常容易�؜淆，如：
>alter database drop logfile group 1;
>alter system switch logfile;
>alter system kill session 'sid,serial#';

��Z��方便记忆�Q�归�U�_��下：
alter database (改变数据�?-database)
和物理文件直接相关的操作

alter system (改变实例--instance)
不直接牵涉到��盘文�g的操�?/font>

当然�Q�这只是��Z��方便记忆�Q�只满��大部分的情况�Q��ƈ不是所有的命��o都能套到�q�个框框里，最�l�还是用熟了��p��住哪个是哪个了�?/font>

---外传---
oracle server=Instance (memory structure) + Database (disk file) + U/S Process
   Instance=SGA + Background Process
   Database=control file + data file + log file
   U/S Process= User Process + Server Process + PGA

Jcat 2009-06-16 12:32 发表评论

RMAN基础

Jcat — Wed, 13 May 2009 07:47:00 GMT
��d��RMAN
[oracle@dcm ~]$ rman target /
Recovery Manager: Release 11.1.0.6.0 - Production on Wed May 13 13:25:30 2009
Copyright (c) 1982, 2007, Oracle. All rights reserved.
connected to target database: O11G (DBID=140043054)
�?/font>
[oracle@dcm ~]$ rman
Recovery Manager: Release 11.1.0.6.0 - Production on Thu Jun 18 13:28:07 2009
Copyright (c) 1982, 2007, Oracle. All rights reserved.
RMAN> connect target /
connected to target database: O11G (DBID=140043054)

因�ؓ数据库是Open的，且又是非归��模式�Q�所以无法进行在�U�全备䆾
RMAN> backup database;
ORA-19602: cannot backup or copy active file in NOARCHIVELOG mode

关了�Q�整成mount状�?/font>
RMAN> shutdown
database closed
database dismounted
Oracle instance shut down

RMAN> startup mount
connected to target database (not started)
Oracle instance started
database mounted

开始数据库全备�?/font>
RMAN> backup database;
Starting backup at 13-MAY-09
allocated channel: ORA_DISK_1
channel ORA_DISK_1: SID=154 device type=DISK
channel ORA_DISK_1: starting full datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
input datafile file number=00001 name=/u01/app/oracle/oradata/o11g/system01.dbf
input datafile file number=00002 name=/u01/app/oracle/oradata/o11g/sysaux01.dbf
input datafile file number=00005 name=/u01/app/oracle/oradata/o11g/example01.dbf
input datafile file number=00003 name=/u01/app/oracle/oradata/o11g/undotbs01.dbf
input datafile file number=00007 name=/u01/app/oracle/oradata/o11g/testspace2.dbf
input datafile file number=00004 name=/u01/app/oracle/oradata/o11g/users01.dbf
channel ORA_DISK_1: starting piece 1 at 13-MAY-09
channel ORA_DISK_1: finished piece 1 at 13-MAY-09
piece handle=/u01/app/oracle/flash_recovery_area/O11G/backupset/2009_05_13/o1_mf_nnndf_TAG20090513T153229_50nxwnqz_.bkp tag=TAG20090513T153229 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:01:35
channel ORA_DISK_1: starting full datafile backup set
channel ORA_DISK_1: specifying datafile(s) in backup set
input datafile file number=00006 name=/u01/app/oracle/oradata/o11g/testspace.dbf
channel ORA_DISK_1: starting piece 1 at 13-MAY-09
channel ORA_DISK_1: finished piece 1 at 13-MAY-09
piece handle=/u01/app/oracle/flash_recovery_area/O11G/backupset/2009_05_13/o1_mf_nnndf_TAG20090513T153229_50nxzv00_.bkp tag=TAG20090513T153229 comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:02
Finished backup at 13-MAY-09

Starting Control File and SPFILE Autobackup at 13-MAY-09
piece handle=/home/oracle/myo11g/autobackup/c-140043054-20090513-00 comment=NONE
Finished Control File and SPFILE Autobackup at 13-MAY-09

查看备䆾信息
RMAN> list backup of database;
List of Backup Sets
===================
BS Key Type LV Size       Device Type Elapsed Time Completion Time
------- ---- -- ---------- ----------- ------------ ---------------
1       Full    1.15G      DISK        00:01:33     13-MAY-09
        BP Key: 1   Status: AVAILABLE Compressed: NO Tag: TAG20090513T153229
        Piece Name: /u01/app/oracle/flash_recovery_area/O11G/backupset/2009_05_13/o1_mf_nnndf_TAG20090513T153229_50nxwnqz_.bkp
List of Datafiles in backup set 1
File LV Type Ckp SCN    Ckp Time Name
---- -- ---- ---------- --------- ----
1       Full 1565621    13-MAY-09 /u01/app/oracle/oradata/o11g/system01.dbf
2       Full 1565621    13-MAY-09 /u01/app/oracle/oradata/o11g/sysaux01.dbf
3       Full 1565621    13-MAY-09 /u01/app/oracle/oradata/o11g/undotbs01.dbf
4       Full 1565621    13-MAY-09 /u01/app/oracle/oradata/o11g/users01.dbf
5       Full 1565621    13-MAY-09 /u01/app/oracle/oradata/o11g/example01.dbf
7       Full 1565621    13-MAY-09 /u01/app/oracle/oradata/o11g/testspace2.dbf

BS Key Type LV Size       Device Type Elapsed Time Completion Time
------- ---- -- ---------- ----------- ------------ ---------------
2       Full    252.00K    DISK        00:00:07     13-MAY-09
        BP Key: 2   Status: AVAILABLE Compressed: NO Tag: TAG20090513T153229
        Piece Name: /u01/app/oracle/flash_recovery_area/O11G/backupset/2009_05_13/o1_mf_nnndf_TAG20090513T153229_50nxzv00_.bkp
List of Datafiles in backup set 2
File LV Type Ckp SCN    Ckp Time Name
---- -- ---- ---------- --------- ----
6       Full 1565621    13-MAY-09 /u01/app/oracle/oradata/o11g/testspace.dbf

RMAN> list backup of controlfile;
List of Backup Sets
===================
BS Key Type LV Size       Device Type Elapsed Time Completion Time
------- ---- -- ---------- ----------- ------------ ---------------
3       Full    9.36M      DISK        00:00:08     13-MAY-09
        BP Key: 6   Status: AVAILABLE Compressed: NO Tag: TAG20090513T161909
        Piece Name: /home/oracle/myo11g/autobackup/c-140043054-20090513-01
Control File Included: Ckp SCN: 1568079      Ckp time: 13-MAY-09

删除指定(BS Key)的备份集
RMAN> delete backupset 1;
using channel ORA_DISK_1
List of Backup Pieces
BP Key BS Key Pc# Cp# Status      Device Type Piece Name
------- ------- --- --- ----------- ----------- ----------
1       1       1   1   AVAILABLE   DISK        /u01/app/oracle/flash_recovery_area/O11G/backupset/2009_05_13/o1_mf_nnndf_TAG20090513T153229_50nxwnqz_.bkp

Do you really want to delete the above objects (enter YES or NO)? YES
deleted backup piece
backup piece handle=/u01/app/oracle/flash_recovery_area/O11G/backupset/2009_05_13/o1_mf_nnndf_TAG20090513T153229_50nxwnqz_.bkp RECID=1 STAMP=686763156
Deleted 1 objects

删除所有备份集
RMAN> delete backup;

using channel ORA_DISK_1

List of Backup Pieces
BP Key BS Key Pc# Cp# Status      Device Type Piece Name
------- ------- --- --- ----------- ----------- ----------
2       2       1   1   AVAILABLE   DISK        /u01/app/oracle/flash_recovery_area/O11G/backupset/2009_05_13/o1_mf_nnndf_TAG20090513T153229_50nxzv00_.bkp
3       3       1   1   AVAILABLE   DISK        /home/oracle/myo11g/autobackup/c-140043054-20090513-00

Do you really want to delete the above objects (enter YES or NO)? YES
deleted backup piece
backup piece handle=/u01/app/oracle/flash_recovery_area/O11G/backupset/2009_05_13/o1_mf_nnndf_TAG20090513T153229_50nxzv00_.bkp RECID=2 STAMP=686763258
deleted backup piece
backup piece handle=/home/oracle/myo11g/autobackup/c-140043054-20090513-00 RECID=3 STAMP=686763268
Deleted 2 objects

Jcat 2009-05-13 15:47 发表评论

Oracle里的�Ҏ��数字

Jcat — Thu, 23 Apr 2009 12:33:00 GMT
01555
ORA-01555: snapshot too old   回滚�D�不够用�?br />可以采取的措施有�Q?br />1、应用程序尽量避免巨表的漫长查询操作�Q�改传统的cursor游标为bulk collect�Q?br />2、尽量程序中不要使用大事务量的增删改操作�Q�同时记得及时commit�Q?br />3、加大undo表空间和加大undo的retention�?br />
10046
Event 10046是oracle用于�pȝ��性能分析时的一个最重要的事件�?/font>

Jcat 2009-04-23 20:33 发表评论

Oracle buys Sun

Jcat — Tue, 21 Apr 2009 02:27:00 GMT
Ora s le

74亿美�?/span>呀�Q�现在Oracle可以�?8摸全面抗衡了�Q?br />
��直就成了IT届的 GE vs 襉K��?

----收购后的格局----
航母�U�：IBM vs Oracle+SUN
软�g�Q�Microsoft�Q�SAP�Q�RedHat�Q�Sybase�Q�小赛的�U�别是不够放在这的，我觉得它最�l�也会走上被收购的道路）
��g�Q�HP�Q�Dell�Q�Fujitsu

----预测----
1. MySQL是生是死�Q?br />Oracle是当今企业��数据库的老大�Q�MySQL是当今互联网应用的老大�?br />如果��O把小My�q�掉�Q��ƈ不代表小My的那部分市场会成为小O的；相反�Q�如果小O能好好照��ְ�My�Q�那��直就无敌了�?br />个�h预测�Q?5分，只要��O能够保持一颗开攄��心，��My应该能有很好的发展�?br />
2. Solaris是生是死�Q?br />��O一直致力于发展Linux技术（比如他和��红合作搞的Unbreakable Linux�Q�，是把Solaris拿来当补充，�q�是�q�掉�Q?br />个�h预测�Q?0分，不会有太大发展，保持现状�?br />
3. Java的发展方向？
Oracle在很多方面都很需要Java�Q�这是好的一面�?br />但是Java作�ؓ一门技术，而不是一个��品，需要有一个中立的代表�Q�SUN以前的态度�q�是不错的。但��O和小I昄��是誓不两立的�Q�会不会��D��Java世界的分裂？
个�h预测�Q?0分，�q�个可能最��N��，实在不希望看��C��好的�l�果�?br />
4. ��g部分
�q�是��O没有的，但是不排除小O把它剥离卖出厅R��但我又觉得�q�种可能性很��，因�ؓ��O最不缺的就是钱�Q�而且�q�是��O��d��I叫板的一大资本�?br />个�h预测�Q?0分，以后Oracle也可以玩total solution的游戏了�?br />

----关系----
和IBM�Q�这回真的成为死��d��头了
和HP、Dell�Q�选什么��品还是由市场军_��的，��O��M��能强买强卖SUN的服务器吧，�q�好�?br /> 和RedHat�Q�小O和小�U�正在一��h��q�Unbreakable Linux�Q�我觉得��O会��l�搞下去
和Sybase�Q�这回搞得三大主��操作系�l�（Windows、AIX、Solaris�Q�都有自��q��数据库了�Q�Sybase的数据库��越来越隑֍��?br /> 和SAP�Q�Microsoft�Q�跟�q�次收购没太大关�p�，但震��g��定不��?br />

Jcat 2009-04-21 10:27 发表评论

��试Oracle Job

Jcat — Thu, 16 Apr 2009 07:42:00 GMT
--最多同时运行的JOB个数�Q�如果太��，JOB��排队等待；如果�?�Q�就没有JOB会被执行�?/font>
SQL> show parameter job_queue_processes
NAME                                        TYPE          VALUE
--------------------------         ----------    -------------
job_queue_processes         integer          10

--一个什么也不做的SP
CREATE OR REPLACE PROCEDURE mytest
IS
BEGIN
   NULL;
END;

--一个往表里写数据的SP
CREATE OR REPLACE PROCEDURE mytest
IS
BEGIN
   insert into test_table values(...);
END;

--定时调用SP�Q?0�U�（86400�U?1天）
SQL> VARIABLE job_id NUMBER;
SQL> BEGIN
           -- :job_id中的冒号表示冒号后面的是变量�Q�类似SQL Server的@
           DBMS_JOB.SUBMIT(:job_id, 'mytest;', sysdate, 'sysdate + 10 / 86400') ;
           COMMIT; --记得一定要commit�?/font>
           END;

--查看刚才生成的Job ID
SQL> set serveroutput on
SQL> execute dbms_output.put_line(:job_id) --�q�里倒是不需要commit�Q�直接execute��好�?/font>
318

--查看所有Job
select * from user_jobs;

--删除Job
SQL> BEGIN
           DBMS_JOB.REMOVE(123); --123是Job ID
           COMMIT;
           END;

如果Job�׃��某种原因未能成功�q�行�Q�Oracle��重�?6�ơ，之后如果�q�未能成功运行，��被标记为Broken�?br />

http://www.tkk7.com/Jcat/archive/2009/12/17/306315.html
�?0g开始，DBMS_SCHEDULER 逐步会替换掉 DBMS_JOB

DBMS_JOB has been around forever, and now it is deprecated. Although DBMS_JOB still exists in 10g and 11g, but only for backward compatibility. No new features are being added to dbms_job and you will likely quickly run into its limitations. Oracle recommends the use of DBMS_SCHEDULER in releases 10g and up. DBMS_SCHEDULER is a much more robust package and fully-featured than DBMS_JOB. To use the DBMS_SCHEDULER package a user must be granted the CREATE JOB privilege.

After replace DBMS_JOB with DBMS_SCHEDULER for all jobs successful, the job_queue_processes parameter can now be set to zero.
SQL> alter system set job_queue_processes=0;

Jcat 2009-04-16 15:42 发表评论

MySQL CMD

Jcat — Fri, 27 Mar 2009 14:33:00 GMT
��d�� shell> mysql -u用户�?-p密码

查看数据�?/b> mysql> show databases;
臛_��会显�C�出两个数据库mysql和test�Q�这是系�l�自建的�Q�供大家�l�习用�?/font>

使用数据�?/b> mysql> use 数据库名

查看�?/b> mysql> show tables;

查看表结�?/b> mysql> desc 表名;

备䆾数据�?/b> shell> mysqldump -uroot -p密码数据库名 > 备䆾的文件�\�?br />

Jcat 2009-03-27 22:33 发表评论

(Z) �?SQL_Trace 诊断Oracle中的问题

Jcat — Fri, 19 Dec 2008 07:19:00 GMT
很多时候，在我们进行数据库操作�Ӟ��比如drop user,drop table�{�，�l�常会遇到这��L��错误�Q?br />ORA-00604: error occurred at recursive SQL level 1 .

关于Recursive SQL错误�Q?br />我们知道�Q�当我们发出一条简单的命��o以后�Q�Oracle数据库要在后台解析这条命令，�q��{换�ؓOracle数据库的一�p�d��后台操作�Q�这些后台操作统�U�Cؓ递归SQL�?br />比如create table�q�样一条简单的DDL命��o�Q�Oracle数据库在后台�Q�实际上要把�q�个命��o转换为对于obj$�Q�tab$�Q�col$�{�底层表的插入操作。Oracle所作的工作可能比我们有时候想的要复杂的多.�?br />
所以对于这��L��提示�Q�很多时候是没有丝毫用处的�?br />
�q�时�Q�我们需要更详细的信息，sql_trace��是一个好办法�?br />sql_trace实际上就是Oracle内部事�g10046的第一个��别，即level=1。通过event 10046, 我们可以得到详细的数据库会话的信息，每个会话�I�竟在干什么，在等什么。通过对这个事件的学习�Q�我们可以深入的理解和调整Oracle�?br />

�l�合eygle的文章整理的�Q�SQL Trace更详�l�的用法误��原文�?/font>

----------

Event 10046是oracle用于�pȝ��性能分析时的一个最重要的事件。当�Ȁ�z�这个事件后�Q�将通知oracle kernel�q�踪会话的相兛_��时信息，�q�写入到相应trace文�g中。这些有用的信息主要包括sql是如何进行解析，�l�定变量的��用情况，话中发生的等待事件等�?br />Trace文�g��会存放�?font color="#ff0000"> user_dump_dest 所指向的目录，如，C:\ent\oracle\admin\o10g01\udump

Event 10046可分成不同的�U�别�Q�level�Q�，分别�q�踪记录不同�E�度的有用信息。对于这些不同的�U�别�Q�应当注意的是向下兼容的�Q�即高一�U�的trace信息包含低于此��的所有信息�?br />    level 1�Q�跟�t�sql语句�Q�包括解析、执行、提取、提交和回滚�{��?br />    level 4�Q�包括变量的详细信息
    level 8�Q�包括等待事�?br />    level 12�Q�包括绑定变量与�{�待事�g
其中�Q�level 1相当于打开了sql_trace�?br />
在当前会话启用event�Q�可以利用alter session + 事�g名称 + level�Q?br />SQL> alter session set event �?0046 trace name context forever, level 12�?

关闭当前会话的event�Q?br />SQL> alter session set event �?0046 trace name context off�?

�q�里应当值得一提的是，TRACE��消耗相当的�pȝ��资源�Q�因此我们在使用TRACE的时候应当慎重。对于正式的�pȝ��Q�应当只在必要的时候进行TRACE操作�Q��ƈ且应当及时关闭�?br />

Jcat 2008-12-19 15:19 发表评论

The meaning of Oracle's version number

Jcat — Mon, 17 Nov 2008 14:32:00 GMT
Oracle Database的大版本��L��两部分组成：数字+字母�Q�数字自然是大版本号�Q?b>字母则是代表了这个版本的“中心思想�?/b>

8i�Q?998�Q�，9i�Q?001�Q�：那时候正好互联网兴�v�Q�i��׃��表internet�Q�网�l�）
10g�Q?003�Q�，11g�Q?007�Q�：现在��L��分布式计��，g��׃��表grid�Q�网��|��

另外�Q�还有个11i�Q�这�Ҏ��引�v��h��Q�其�?1i是指Oracle EBS�Q�E-Business Suite电子商务套�g�Q�的版本��P��其实��是Oracle ERP。（�q�好像跟Oracle Database没什么必然联�p�，我也不知道这个i代表什么）

�q�也说明Oracle不等于Oracle Database。��^常我们说“Oracle”，一般都是指Oracle Database�Q�当然这是它最著名的��品；不过�Q�其实Oracle早就已经不是一家数据库厂商了，而是航母�U�厂商，东收西购�Q�啥都有了�?/font>

我见�q�这��L��问题�Q��?/font>从版本命名来�?i-9i-10g-11i�Q�是不是意味着grid只是昙花一玎ͼ�”显然这��是��h��了Oracle Database和Oracle ERP�?/font>

----以下是一些参�?---
In 1998, Oracle announced Oracle8i, which is sometimes referred to as Version 8.1 of the Oracle8 database. The i was added to denote added functionality supporting Internet deployment in the new version. Oracle9i followed, with Application Server available in 2000 and Database Server in 2001.

Oracle Database 10g was introduced in 2003; the g denotes Oracle's focus on emerging grid deployment models.

Oracle 11i refers to the Oracle ERP Application Suite (often called Oracle E-Business Suite) and really has nothing to do with the database. The latest release (12) will be referred to as 12i. The confusing part is that Oracle used to call its database "i" for "internet" in releases 8 and 9, but now calls the database "g" for "grid". So, to summarize, Oracle 11i refers to the apps and Oracle 11g refers to the database

Jcat 2008-11-17 22:32 发表评论

Hash Hash

Jcat — Fri, 18 Jul 2008 08:55:00 GMT

Hash�Q�一般翻译做“散列”，也有直接韌��为“哈希”的�?br />
我们通常说的Hash�Q�其实指的是Hash��法�Q�即散列��法�Q�：把�Q意长度的输入�Q�通过Hash��法�Q�变换成固定长度的输出（即Hash倹{��散列��|��?br />
Hash��法有多�U�实现�Ş式，比如MD5、SHA1�Q�这里就不对��法本��n�q�行讨论了。重点谈谈Hash��法的特性和作用吧：
    1�Q�Hash是一�U�压�~�映��：散列值的长度通常�q�小于输入的长度�Q�比如emule�Q��Q意大��是视频文�g�Q�都可以映射成一个固定长度的字符�Ԍ��它就相当于这个文件的信息摘要�Q?br />    2�Q�不同的输入可能会Hash成相同的输出�Q�这是一个小概率事�g�Q�且采用安全性高的Hash��法�Ӟ��两个不同的输入几乎不可能得到相同的Hash�l�果�?br />    3�Q�与加密��法不同�Q�Hash��法是一个不可逆的单向函数�Q�不可能从散列值来唯一的确定输入倹{�?br />

0. 在下面的讨论中，我们��用��C��国这个例子：假设我们要做个存储结构，需要存储下来三国中的英雄，以及他们的详�l�信息。我们用他们的名字来作�ؓ存储的关键��|��例如�Q�刘备，关羽�Q�张飞，�Ҏ��Q�孙�?..n�?br />
1. 一般的�?/b>
    name   age   �w�高   体重
    刘备   30   175   60
    关羽   28   190   80
    张飞   27   185   80
    ...n
�q�时�Q�如果我们想要查找某个英雄，��需要一辚w��历表�Q�一边对名字�q�行比较。它的时间复杂度为O�Q�n�Q?-�U�性阶�?br />
�ȝ��Q�记录在�l�构中的相对位置和记录的关键字之间不存在��定的关�p�，在结构中查找记录旉��q�行一�p�d��和关键字的比较，查找的效率与比较�ơ数密切相关�?br />
2. 数组
    1�Q�通过地址的访问方式（数组�Q�：Array[2]�Q�这�?��是它的“物理”地址�Q�找�?所对应的内�Ҏ��通过“直接定位”，而不是“遍历比较”。所以，在所有的�U�性数据结构中�Q�数�l�的定位速度最快。它的时间复杂度为O�Q?�Q?-常熟阶�?br />
    �q�时�Q�我们会惻I��那我们把�?�?�?...n�Q�按�?-2-3-...n的顺序放在一个数�l�里�Q�数�l�的内容��是�q�个英雄的对象）不就可以直接定位了？
    arrayid   �Q�name   age   �w�高   体重�Q?br />    1            �Q�刘备 �? 30   175   60�Q?br />    2            �Q�关��q��? 28   190   80�Q?br />    3            �Q�张飞 �? 27   185   80�Q?br />    ...n

    2�Q�但是：
        a�Q�可你凭什么说�?1�Q�关=2呢？�Q�显然不是叫你用拼音或者笔��L��序）
        b�Q�我们查扄��时候，不可能用arrayid作�ؓ查找条�g�Q�那�q�查个P呀�Q�，怎样通过name=关羽�Q�“计��”得到arrayid=2呢？

3. 散列�?/b>�Q�其主要目的是用于解��x��据的快速定位问题，也即�Q?font color="#ff1493">怎样通过关键字，计算�Q�而不是遍历比较）得到物理地址�?/font>

1�Q�存放时
    Hash�Q�刘备）=13
    Hash�Q�关��）=7
    Hash�Q�张飞）=26
    ...n

    arrayid   �Q�name   age   �w�高   体重�Q?br />    ...
    7            �Q�关��q��? 28   190   80�Q?br />    ...
    13          �Q�刘备 �? 30   175   60�Q?br />    ...
    26          �Q�张飞 �? 27   185   80�Q?br />    ...n

    我们也可以看出：
    a�Q?font color="#ff0000">散列�Q�数据�ƈ非紧凑的、按��序存入的：我们可以先在13的位�|�存上刘备，再在7的位�|�存上关��；有些位置�q�可能一直是�I�的�?br />    b�Q?font color="#ff0000">�I�间换时�?/font>�Q�因为散列，昄��数组的大��要大于实际实际数据的数量。比如，即��我们只存刘关�?个�h�Q�我们也需要一个至��?6的数�l��?br />
2�Q�查找时�Q�就很简单了
    Aarry[Hash�Q�关��）]=Array[7]

3�Q�当然实际的hash��法�Q�要比上�q�复杂，比如
    “把Key通过Hash��法转换成一个整型数字，然后��将该数字对数组长度�q�行取余�Q�取余结果就当作数组的下标，��value存储在以该数字�ؓ下标的数�l�空间里�Q?br />    而当使用哈希表进行查询的时候，��是再次使用哈希函数��key转换为对应的数组下标�Q��ƈ定位到该�I�间获取value�Q�如此一来，��可以充分利用到数组的定位性能�q�行数据定位。�?br />
4�Q�Hash冲突�Q�前面说了，不同的输入是有可能hash成相同的输出�?br />    如果Hash�Q�关��）=Hash�Q�张飞）=250�Q�我们岂不是在存储过�E�中会遇到麻烦，怎么安排他们二位的地方呢?�Q��M��能让二位打一�Ӟ��谁赢了谁呆在那吧�Q�，�q�就需要一个解军_��H�的�Ҏ��?br />    先存储好了关��，当张飞进入系�l�时会发现关��已�l�是250了，那咱��加一位，251得了。我们查扑ּ�飞的时候也是，一�?50不是张飞�Q�那��加�?�Q�就扑ֈ�了�?br />    更极端地�Q�如果这时Hash�Q��n云）=251�Q�张飞已�l�早早占了他的地方，那就再加1存到252呗。呵呵，�q�时我们会发玎ͼ�当哈希函数冲�H�发生的机率很高�Ӟ��可能会有一��英雄在250�q�个值后面扎堆排队。要命的是查扄��时候，旉��法复杂度早已不是O�Q?�Q�了。所以我们说理想情况下哈希表的时间算法复杂度为O�Q?�Q��?br />    �q�就是说哈希函数的编写是哈希表的一个关键问题，会涉及到一个存储值在哈希表中的统计分布。如果哈希函数已�l�定义好了，冲突的解军_��成�ؓ了改变系�l�性能的关键因素。其实还有很多种�Ҏ��来解军_��H�情况下的存储和查找问题�Q�不一定非要线性向后排队，如果有好的哈希表冲突的解��x��法也能很大程度上提高�pȝ��的效率�?br />

FYI�Q�比较书面化的定义：
    理想的情冉|��能直接找到需要的记录�Q�因此必��d��记录的存储位�|�和它的关键字之间徏立一��定的对应关�p�f�Q��每个关键字和�l�构中一个唯一的存储位�|�相对应。因而查找时�Q�只需�Ҏ��q�个对应关系f扑ֈ��l�定值K的像f�Q�K�Q�。若�l�构中存在关键字和K相等的记录，则必定在f�Q�K�Q�的存储位置上，由此不需要进行比较便可直接取得所查记录。在此，�U�这个对应关�p�f为哈希函敎ͼ�按这个思想建立的表为哈希表�Q�又�U�Cؓ杂凑法或散列法）�?/font>

参考：
http://calmness.javaeye.com/blog/184465
http://blog.163.com/xx_snoopy/blog/static/12577162008426731299/

Jcat 2008-07-18 16:55 发表评论

Jcat — Wed, 16 Apr 2008 06:16:00 GMT
1NF�Q�属性单一�Q�不可再分（仅一个��g��一�U��|��
很显�Ӟ��在当前的��M��RDBMS中，�ȝ��也不可能做出不符合第一范式的数据库�Q�因为RDBMS不允�怽�把数据库表的一列再分成二列或多列。实际上�Q�还真有�q�样的垃圾设计，我在某个��目中遇刎ͼ��?first_name�Q�middle_name�Q�last_name攑֜�一个字�D늚��Q�最后想分都分不开�Q�完全折服�?br />

2NF�Q�消除了部分依赖�Q�部分依赖由复合主键��D��Q�显然所有单主键的表都符�?NF�Q?br /><反例> 一个表
�Q�学��P�� 评��名称�Q�→�Q�姓名， �q�龄�Q?成�W�Q?学分�Q?br /><部分依赖>
�Q�课�E�名�U�ͼ�→（学分�Q?br />�Q�学��P��→（姓名�Q?�q�龄�Q?br /><�Ҏ��> 改�ؓ三个�?br />�Q�学��P��→（姓名�Q?�q�龄�Q?br />�Q�课�E�名�U�ͼ�→（学分�Q?br />�Q�学��P��评��名称�Q�→�Q�成�l�）

3NF�Q�消除了传递依�?br /><反例>
�Q�学��P��→（姓名�Q?�q�龄�Q?学院�Q?学院地点�Q?学院电话�Q?br /><传递依�?gt;
�Q�学��P��→（学院�Q�→�Q�学院地点，学院电话�Q?br /><�Ҏ��> 改�ؓ两个�?br />�Q�学��P��→（姓名�Q�年龄，学院�Q?br />�Q�学院）→（地点�Q�电话）

Jcat 2008-04-16 14:16 发表评论

(z) 存储定w��

Jcat — Mon, 07 Apr 2008 03:41:00 GMT
存储定w��Q�是该存储设备上可以存储数据的最大数量，通常使用千字节（kb kilobyte�Q�、兆字节�Q�MB megabyte�Q�、吉字节�Q�GB, gigabyte�Q�、太字节�Q�TB �Q�terabyte�Q�和PB(Petabyte)、EB(Exabyte)�{�来衡量�?br />
1KB=2(10)B=1024B�Q?括号中的数字�?的指�?卛_��次�?
1MB=2(10)KB=1024KB=2(20)B�Q?br />1GB=2(10)MB=1024MB=2(30)B�?br />1TB=2(10) GB=1024GB=2(40)B
1PB=2(10) TB=1024TB=2(50)B
1EB=2(10) PB=1024PB=2(60)B

1Byte相當��g��個英文字母�?br />Kilobyte(KB)=1024B相當��g��則短��故事的內容�?br />Megabyte(MB)=l024KB相當��g��則短��小說的文字內容�?br />Gigabyte(GB)=1024MB相當��D��多芬�W�五樂章交響曲的樂譜內容�?br />Terabyte(TB)=1024GB相當��g��家大型醫院中所有的X光圖片資�a�量�?br />Petabyte(PB)=l024TB相當�?0%的全��學術研�I�圖曔R��藏書資訊內容�?br />Exabyte (EB)=1024PB�Q?EB相當��D��今全世界人類所��過的話語�?br />Zettabyte(ZB)=1024EB如同全世界�v灘上的沙子數量總和�?br />Yottabyte(YB)=1024ZB相當�?000位�h��體內的微細胞總和�?br />

厂商刉��硬盘时�Q?1K=1000B�Q?M=1000K�Q?G=1000M
所有标80G的硬盘，拿到电脑里就变小了（除以3�?.024�Q�：74.5 = 80 ÷ 1.07374

-----------

1 Byte = 8 bit�Q?�?�q�制敎ͼ�2�?�ơ方�Q?56

一般网�l�的速度是以“位每秒”（bps�Q�来表示。注意，是小写b�Q�位�Q�，而不是大写B�Q�字节）。所以，昄��的网�l�速度是需要除8�?/font>�?br />比如我的230K的无�U�猫�Q�也��是说它一�U�钟可传�?30K个bit�Q�其实也��?8K个字节，而且�q?8K�q�只是理惛_��|��最快确实达到过20K�Q�一般能�?K都笑了�?img src ="http://www.tkk7.com/Jcat/aggbug/191210.html" width = "1" height = "1" />

Jcat 2008-04-07 11:41 发表评论

Oracle LOB

Jcat — Thu, 27 Mar 2008 03:12:00 GMT
<引言>
    随着�C�会的发展，在现代信息系�l�的开发中�Q�需要存储的已不仅仅是简单的文字信息�Q�同时还包括一些图片和韛_��资料或者是��长的文本。比如开发一套旅�怿�息系�l�，每一个景炚w��有丰富的囄��、音像资料和大量的文字介�l�。这��p��求后台数据库要有存储�q�些数据的能力�?br />    ORACLE自Oracle8i起通过提供LOB字段实现了该功能�?br />

    LOB�Q�即Large Objects�Q�大对象�Q�，是用来存储大量的二进制和文本数据的一�U�数据类型（一个LOB字段可存储可多达4GB的数据）。它又分��Z��U�类型：内部LOB和外部LOB�?br />    1.内部LOB��数据以字节��的形式存储在数据库的内部。因而，内部LOB的许多操作都可以参与事务�Q�也可以像处理普通数据一样对其进行备份和恢复操作。Oracle支持三种�c�d��的内部LOB�Q?br />    1�Q�BLOB�Q�二�q�制数据�Q?br />    2�Q�CLOB�Q�单字节字符数据�Q?br />    3�Q�NCLOB�Q�多字节国家字符数据�Q��?br />    其中CLOB和NCLOB�c�d��适用于存储超长的文本数据�Q�BLOB字段适用于存储大量的二进制数据，如图像、视频、音频等�?br />
    2.目前只支持一�U�外部LOB�c�d��Q�即BFILE�c�d��。在数据库内�Q�该�c�d��仅存储数据在操作�pȝ��中的位置信息�Q�而数据的实体以外部文件的形式存在于操作系�l�的文�g�pȝ��中。因而，该类型所表示的数据是只读的，不参与事务�?br />    该类型可帮助用户��理大量的由外部�E�序讉K��的文件�?br />

    LOB�׃��部分�l�成�Q�数据（��|��和指向数据的指针�Q�定位器�Q�。尽��g��表自�w�一起存储，但是一个LOB列�ƈ不包含��|��仅有它的定位指针。更�q�一步，��Z��使用大对象，�E�序必须声明定位器类型的本地变量�?br />
    当内部LOB被创建时�Q�定位器被存攑֜�列中�Q��D��存放在LOB�D�中�Q�LOB�D�|��在数据库内部表的一部分�?br />    当BFILE被创建时�Q�定位器如同�q�_��一样存储在列中�Q�但是它的��D��存储在数据库之外的操作系�l�文件中�?img src ="http://www.tkk7.com/Jcat/aggbug/188959.html" width = "1" height = "1" />

Jcat 2008-03-27 11:12 发表评论

Sybase PD vs PB

Jcat — Mon, 10 Mar 2008 08:55:00 GMT

PowerDesigner

Enterprise architecture modeling and design solution

Build or re-engineer your applications using PowerDesigner, Sybase's all-in-one enterprise modeling and design solution.

PowerDesigner 12.5, a model-driven approach to aligning business and IT, is an enterprise modeling and design solution that helps you implement effective enterprise architecture. It brings powerful analysis and design techniques to your development lifecycle.
PowerDesigner�Q�数据库建模工具�Q�与ERWin�{�工兯��v名，它的归类�?Modeling & Metadata

PowerBuilder

A 4GL rapid application development tool

Accelerate development with PowerBuilder, the market’s favorite 4GL RAD tool.

PowerBuilder lets you stay ahead of the development curve and offers the flexibility to develop for your most productive platform. PowerBuilder 11.1 incorporates new and emerging technologies that let you to build traditional two-tier applications, distributed applications, Web applications and Smart Clients with speed and ease. And with this new release, you can easily deploy your existing PowerBuilder applications to .NET. Get started with PowerBuilder 11.1 today, and see how quickly and easily you can build data-driven applications.

PowerBuilder�Q�是公司随数据库产品的一个前台开发工��P��专门开发应用程序的。它的归�c�L�� Development

Jcat 2008-03-10 16:55 发表评论

Jcat — Mon, 10 Mar 2008 08:33:00 GMT

Adaptive Server Enterprise

Mission-critical enterprise data management

Adaptive Server Enterprise is the mission-critical data management system for taming the information explosion.

Adaptive Server Enterprise (ASE) has long been noted for its reliability, low total cost of ownership and superior performance. With its latest version, ASE 15, it has been dramatically enhanced to deliver capabilities urgently needed by enterprises today. It lays the long-term foundation for strategic agility and continuing innovation in mission-critical environments. ASE 15 provides unique security options and a host of other new features that boost performance while reducing operational costs and risk. Find out how you can exploit new technologies such as grids and clusters, service-oriented architectures and real-time messaging.
ASE�Q�企业及数据库服务器�Q�适合用于企业�U�OLTP�Q�和混合负蝲�Q�环�?

SQL Anywhere

Sybase iAnywhere - Mobile and Embedded Database

The industry-leading mobile and embedded database, providing data management and exchange technologies that enable applications for frontline environments.

SQL Anywhere provides data management and exchange technologies designed for database-powered applications that operate in frontline environments without onsite IT support. It offers enterprise-caliber features in a database that is easily embedded and widely deployed in server, desktop, remote office and mobile applications.
ASA�Q�Adaptive Server Anywhere�Q�：轻量�U�数据库服务器，适合用于�U�d��计算�Q�PDA、带操作�pȝ��的手机等�Q?嵌入式计��（POS机、�\由器�{�）�Q�和工作�l��OLTP环境�Q�特别是�U�d��计算/嵌入式计��。（我估计ASA现在改名叫SQL Anywhere了）

Sybase IQ

Highly optimized business intelligence, analytics and data warehousing

The smartest way to deliver high-performance enterprise analytics and data warehousing with dramatically faster results at a low cost.

Sybase IQ is a highly optimized analytics server designed specifically to deliver faster results for mission-critical business intelligence, data warehouse and reporting solutions on any standard hardware and operating systems. It works with diverse data �?including unstructured data �?and diverse data sources to deliver unsurpassed query performance at the lowest price/performance available.
ASIQ�Q�企业��数据仓库数据库服务器�Q�适合于DSS环境中的企业�U�数据仓库和数据集市

谨以此篇记录我开始Sybase之旅的第一�?/font>

Jcat 2008-03-10 16:33 发表评论

(z) SQL Server vs Oracle 存储�q�程语法转换

Jcat — Thu, 13 Sep 2007 07:42:00 GMT

1. top N 问题
在sql server中，top N 问题很容易解冻I��如下例：从表stbdbdj中选取排序后的�W�一行数据进行赋倹{�?/span>

在sql中解��x��法很��单，在select 后面加上�Q�top n 卛_��Q�其�?n 代表行数�?

select top 1 @entrust_date = entrust_date,
@entrust_no = entrust_no
from run2k..stbdbdj
where entrust_date = @date
and entrust_no > @entrust_no_q
and report_status = '1'
order by entrust_date,entrust_no�Q?/span>

在oracle中，没有top n�q�个命��o�Q�我们采取把两层查询方式解决�Q�首先，把需要查扄��字段值直接进行排序，然后在外面进行第二次查询�Q��ƈ使用rownum军_��行数�?br />
select entrust_date,entrust_no
into @entrust_date, @entrust_no
from ( select entrust_date,entrust_no
from stbdbdj
where entrust_date = @date
and entrust_no > @entrust_no_q
and report_status = '1'
order by entrust_date,entrust_no )
where rownumber <=1 ;

2. 如何解决�l�果集返回时�Q? 和变量同时存在的问题
下面例子表示�Q�在用游标返回结果集�Ӟ��同时�q�回一个变量的��|��?/span> sql server 中代码如下所�C�：
select a.*,b.organ_id
from run2k..stbbp a,run2k..stkaccoarg b
where a.date = @entrust_date
and a.serial_no = @serial_no
and a.branch_no = b.branch_no
and a.exchange_type = b.exchange_type;

但在oracle中却没有�q�种用法�Q��?’后面必需跟from。解��x��法如下：
1�Q�我们可以把 '*' 变成所需要选择的字�D�，��是说采用表中需要显�C�的全部字段表示*�?br /> 例如�Q?br />

open p_cursor for
select branch_no,...,organ_id
where ...

2�Q�如果这个字�D�|��者说变量是从另外一张表中取出来的，同样可以采用下面的办法�?br />
open p_cursor for
select a.*,b.organ_id;
from stkaccoentrust a, stkaccoarg b
where a.branch_no = b.branch_no
and a.exchange_type = b.exchange_type
and a.init_date = v_entrust_date
and a.serial_no = v_serial_no;

3. 外联接问�?/b>
sql <---> oracle
a = *b <---> a(+)= b
a *= b <---> a = b(+)

4. 多条记录求和问题
select sum(A+B+C)
into D
from ...
where ...
group by ...

单条记录求和
select A+B
into C
from ...
where ...

5. case 问题转换
sql:
case client_status
when '0' then '正常'
when '1' then '�ȝ��'
when '2' then '挂失'
when '3' then '销�?
else '未知'
end

oracle:
decode(client_status,'0','正常�Q?1','�ȝ��','2','挂失','3','销�?,'未知');

6. char �?varchar �c�d��区别�Q?
char ��N��补空��|��varchar ��N��不补�I�格�?b>

7. convert转换问题
sql ---> oracle
convert(char(5),branch_no) ---> to_char(branch_no,'99999')
convert(char(19),count(*)) ---> lpad(to_char(count(*)),19)
convert(varchar(20),serial_no) ---> to_char(serial_no,'999...9' )
��d��20�?
lpad(to_char(serial_no),20)

8. charindex(substring,string) ---> instr(string,substring)
子串父串 ---> 父串子串

Jcat 2007-09-13 15:42 发表评论

亚洲AV永久无码天堂影院,亚洲va久久久噜噜噜久久男同 ,国产亚洲综合色就色

Win - 启动�l�色版MYSQL

DB2基础命��o

数据库、实�?on Oracle/DB2

物化视图

IPC extproc PLSExtPro

Oracle Scheduler 10G NEW

数据库的隔离�U�别

一致和�q�发是对立的�Q�需要根据应用，选择权宜之计

NLS_DATE_FORMAT

Oracle Data Dictionary

ALTER DATABASE vs ALTER SYSTEM

RMAN基础

Oracle里的�Ҏ��数字

Oracle buys Sun

���试Oracle Job

MySQL CMD

(Z) �?SQL_Trace 诊断Oracle中的问题

The meaning of Oracle's version number

Hash Hash

(z) 存储定w��

Oracle LOB

Sybase PD vs PB

PowerDesigner

Enterprise architecture modeling and design solution

PowerBuilder

A 4GL rapid application development tool

Adaptive Server Enterprise

Mission-critical enterprise data management

SQL Anywhere

Sybase iAnywhere - Mobile and Embedded Database

Sybase IQ

Highly optimized business intelligence, analytics and data warehousing

(z) SQL Server vs Oracle 存储�q�程语法转换

��试Oracle Job