原則一：注意WHERE子句中的連接順序：
ORACLE采用自下而上的順序解析WHERE子句,根據(jù)這個原理,表之間的連接必須寫在其他WHERE條件之前, 那些可以過濾掉最大數(shù)量記錄的條件必須寫在WHERE子句的末尾.
尤其是“主鍵ID=？”這樣的條件。

原則二： SELECT子句中避免使用 ‘ * ‘：
ORACLE在解析的過程中, 會將'*' 依次轉(zhuǎn)換成所有的列名, 這個工作是通過查詢數(shù)據(jù)字典完成的, 這意味著將耗費(fèi)更多的時間 。

簡單地講，語句執(zhí)行的時間越短越好（尤其對于系統(tǒng)的終端用戶來說）。而對于查詢語句，由于全表掃描讀取的數(shù)據(jù)多，尤其是對于大型表不僅查詢速度慢，而且對磁盤IO造成大的壓力，通常都要避免，而避免的方式通常是使用索引Index。

使用索引的優(yōu)勢與代價(jià)。
優(yōu)勢：
1）索引是表的一個概念部 分,用來提高檢索數(shù)據(jù)的效率，ORACLE使用了一個復(fù)雜的自平衡B-tree結(jié)構(gòu). 通常,通過索引查詢數(shù)據(jù)比全表掃描要快. 當(dāng)ORACLE找出執(zhí)行查詢和Update語句的最佳路徑時, ORACLE優(yōu)化器將使用索引. 同樣在聯(lián)結(jié)多個表時使用索引也可以提高效率.
2） 另一個使用索引的好處是,它提供了主鍵(primary key)的唯一性驗(yàn)證.。那些LONG或LONG RAW數(shù)據(jù)類型, 你可以索引幾乎所有的列. 通常, 在大型表中使用索引特別有效. 當(dāng)然,你也會發(fā)現(xiàn), 在掃描小表時,使用索引同樣能提高效率.
代價(jià)： 雖然使用索引能得到查詢效率的提高,但是我們也必須注意到它的代價(jià). 索引需要空間來存儲,也需要定期維護(hù), 每當(dāng)有記錄在表中增減或索引列被修改時, 索引本身也會被修改. 這意味著每條記錄的INSERT , DELETE , UPDATE將為此多付出4 , 5 次的磁盤I/O . 因?yàn)樗饕枰~外的存儲空間和處理,那些不必要的索引反而會 使查詢反應(yīng)時間變慢.。而且表越大，影響越嚴(yán)重。

使用索引需要注意的地方：

1、避免在索引列上使用NOT ，　
我們要避免在索引列上使用NOT, NOT會產(chǎn)生在和在索引列上使用函數(shù)相同的影響. 當(dāng)ORACLE”遇到”NOT,他就會停止使用索引轉(zhuǎn)而執(zhí)行全表掃描.

2、避免在索引列上使用計(jì)算．
WHERE子句中，如果索引列是函數(shù)的一部分．優(yōu)化器將不使用索引而使用全表掃描． 舉例:

第一章、      開發(fā)成功的Oracle應(yīng)用程序

? You should do it in a single SQL statement if at all possible.

? If you cannot do it in a single SQL Statement, then do it in PL/SQL.

? If you cannot do it in PL/SQL, try a Java Stored Procedure.

? If you cannot do it in Java, do it in a C external procedure.

? If you cannot do it in a C external routine, you might want to seriously think about why it is you need to do it...

 這段表明我們操作一個過程， 簡單來說，優(yōu)先選擇SQL

SQL中一些注意

1、不要在MTS下運(yùn)行長事務(wù)

2、使用綁定變量

理解并行控制

1、鎖

? Oracle locks data at the row level on modification only. There is no lock escalation to a block

or table level, ever.

? Oracle never locks data just to read it. There are no locks placed on rows of data by simple

reads.

? A writer of data does not block a reader of data. Let me repeat – reads are not blocked by

writes. This is fundamentally different from almost every other database, where reads are

blocked by writes.

? A writer of data is blocked only when another writer of data has already locked the row it was

going after. A reader of data never blocks a writer of data.

所以實(shí)際應(yīng)用時候，注意改變鎖的級別!!!

2、多版本

主要針對寫操作時候，對讀操作不阻塞

具體來說： (R-read  W-write)

R/R    　　 不上鎖，但是如果兩個線程先讀再修改，則需要加SS鎖(for update)

R/W  W/R   多版本控制，不阻塞讀

W/W          使用SX鎖，阻塞任何ＤＭＬ操作

DML鎖分類表

表1 Oracle的TM鎖類型
鎖模式	鎖描述	解釋	SQL操作
0	none
1	NULL	空	Select
2	SS(Row-S)	行級共享鎖，其他對象只能查詢這些數(shù)據(jù)行	Select for update、Lock for update、Lock row share
3	SX(Row-X)	行級排它鎖，在提交前不允許做DML操作	Insert、 Update、Delete、Lock row share
4	S(Share)	共享鎖	Create index、Lock share
5	SSX(S/Row-X)	共享行級排它鎖	Lock share row exclusive
6	X(Exclusive)	排它鎖	Alter table、Drop able、Drop index、Truncate table 、Lock exclusive

一.準(zhǔn)備修改。首先必須用sys連接。

1.修改sys密碼.用其他dba權(quán)限的賬戶登陸后輸入：alter user sys identified by 新密碼;

2.連接剛剛修改的sys（要在登陸狀態(tài)下）,CONNECT SYS／密碼 as sysdba;

二.開始修改:

SQL> shutdown immediate;
SQL> startup mount;
SQL> alter system enable restricted session;
SQL> alter system set job_queue_processes=0;
SQL> alter database open;
SQL> alter database character set internal_use UTF-8;
SQL> shutdown immediate;
SQL> startup
      注意：如果沒有大對象，在使用過程中進(jìn)行語言轉(zhuǎn)換沒有什么影響，具體過程如下（切記設(shè)定的字符集必須是ORACLE支持，不然不能start）

1.使用sys登錄Oracle，利用dbms_xdb修改端口設(shè)置

? SQL> -- Change the HTTP/WEBDAV port from 8080 to 8081
? SQL> call dbms_xdb.cfg_update(updateXML(
??? 2??????? dbms_xdb.cfg_get()
??? 3????? , '/xdbconfig/sysconfig/protocolconfig/httpconfig/http-port/text()'
??? 4????? , 8081))
??? 5? /

? Call completed.

? SQL> -- Change the FTP port from 2100 to 2111
? SQL> call dbms_xdb.cfg_update(updateXML(
??? 2???????? dbms_xdb.cfg_get()
??? 3?????? , '/xdbconfig/sysconfig/protocolconfig/ftpconfig/ftp-port/text()'
??? 4?????? , 2111))
??? 5? /

? Call completed.

? SQL> COMMIT;

? Commit complete.

? SQL> EXEC dbms_xdb.cfg_refresh;

? PL/SQL procedure successfully completed.

2.使用OEM console，選擇數(shù)據(jù)庫，XML Database，Configuration。更改XDB的有關(guān)設(shè)置。

3.去掉數(shù)據(jù)庫的初始化參數(shù)：? dispatchers='(PROTOCOL=TCP) (SERVICE=XDB)'，將會禁止XDB的http和ftp服務(wù)

2.
insert into table_tmp
(select * from table1)

3.
update table_tmp
set (name,password) = (select name,password from table1)

4.
select tab_name from tables
where (tab_name,db_ver) =
( 
    select tab_name,db_ver from tab_columns 
    where version = 604
)

?REF CURSOR分兩種：Strong REF CURSOR 和 Weak REF CURSOR。
? Strong REF CURSOR:指定retrun type，CURSOR變量的類型必須和return type一致。
->>>DECLARE
???TYPE strongcurtyp IS REF CURSOR RETURN emp%ROWTYPE;
???emp_cv strongcurtyp;

?Weak REF CURSOR:不指定return type，能和任何類型的CURSOR變量匹配。
->>>DECLARE
???TYPE weakcurtyp IS REF CURSOR;
???weak_cv weakcurtyp;
???any_cv SYS_REFCURSOR; --使用SYS_REFCURSOR可以使變量any_cv跟任何weak REF CURSOR類型匹配。

?

(1)使用Strong REF CURSOR例子
CREATE OR REPLACE PACKAGE emp_data AS
?TYPE empcurtyp IS REF CURSOR RETURN emp%rowtype; --定義TYPE as Strong REF CURSOR
?PROCEDURE open_emp_cv(emp_cv IN OUT empcurtyp, choice INT); --根據(jù)不同的choice選擇不同的CURSOR
?PROCEDURE retrieve_data(choice INT); --通過調(diào)用procedure open_emp_cv，返回指定的結(jié)果集。
END emp_data;

CREATE OR REPLACE PACKAGE BODY emp_data AS

?------------procedure open_emp_cv-----------------------?
?PROCEDURE open_emp_cv(emp_cv IN OUT empcurtyp, choice INT) IS --emp_cv作為傳入/傳出的CURSOR PARAMETER
?BEGIN
??IF choice = 1 THEN
???OPEN emp_cv FOR
???SELECT * FROM emp WHERE empno < 7800;
??ELSIF choice = 2 THEN
???OPEN emp_cv FOR
???SELECT * FROM emp WHERE SAL < 1000;
??ELSIF choice = 3 THEN
???OPEN emp_cv FOR
???SELECT * FROM emp WHERE ename like 'J%';
??END IF;
?END;
?
?
?-----------procedure retrieve_data-----------------------?
?PROCEDURE retrieve_data(choice INT) IS
??return_cv empcurtyp; --定義傳入open_emp_cv的CURSOR變量
??return_row emp%ROWTYPE;
??invalid_choice EXCEPTION;
?BEGIN
??open_emp_cv(return_cv, choice); --調(diào)用 procedure OPEN_EMP_CV
?
??IF choice = 1 THEN
???DBMS_OUTPUT.PUT_LINE('EMPLOYEES with empno less than 7800');
??ELSIF choice = 2 THEN
???DBMS_OUTPUT.PUT_LINE('EMPLOYEES with salary less than 1000');
??ELSIF choice = 3 THEN
???DBMS_OUTPUT.PUT_LINE('EMPLOYEES with name starts with ''J''');
??ELSE
???RAISE invalid_choice;
??END IF;
?
??LOOP
??FETCH return_cv INTO return_row;
??EXIT WHEN return_cv%NOTFOUND;
??
???DBMS_OUTPUT.PUT_LINE(return_row.empno || '--' || return_row.ename || '--' || return_row.sal);
??END LOOP;
?
?EXCEPTION
??WHEN invalid_choice THEN
???DBMS_OUTPUT.PUT_LINE('The CHOICE should be in one of (1,2,3)!');
??WHEN OTHERS THEN
???DBMS_OUTPUT.PUT_LINE('Errors in procedure RETRIEVE_DATA!');
?END;

END emp_data;

================================
執(zhí)行：
SQL> EXEC emp_data.retrieve_data(1);
EMPLOYEES with empno less than 7800
7369--SMITH--800
7499--ALLEN--1600
7521--WARD--1250
7566--JONES--2975
7654--MARTIN--1250
7698--BLAKE--2850
7782--CLARK--2450
7788--SCOTT--3000
PL/SQL procedure successfully completed

SQL> EXEC emp_data.retrieve_data(2);
EMPLOYEES with salary less than 1000
7369--SMITH--800
7900--JAMES--950
PL/SQL procedure successfully completed

SQL> EXEC emp_data.retrieve_data(3);
EMPLOYEES with name starts with 'J'
7566--JONES--2975
7900--JAMES--950
PL/SQL procedure successfully completed

SQL> EXEC emp_data.retrieve_data(34);
The CHOICE should be in one of (1,2,3)!
PL/SQL procedure successfully completed

(2)使用Weak REF CURSOR例子
-----------procedure open_cv----------------------------
create or replace procedure open_cv(choice IN INT, return_cv OUT SYS_REFCURSOR) is
--參數(shù)return_cv為weak REF CURSOR,利用SYS_CURSOR來定義
begin
?if choice = 1 then
??open return_cv for 'select * from emp';
?elsif choice = 2 then
??open return_cv for 'select * from dept';
?end if;
end open_cv;

-----------procedure retrieve_data----------------------------
create or replace procedure retrieve_data(choice IN INT) is
?emp_rec emp%rowtype;
?dept_rec dept%rowtype;
?return_cv SYS_REFCURSOR;
?invalid_choice exception;
begin
?if choice=1 then
??dbms_output.put_line('employee information');
??open_cv(1,return_cv); --調(diào)用procedure open_cv;
??loop
??fetch return_cv into emp_rec;
??exit when return_cv%notfound;
???dbms_output.put_line(emp_rec.empno||'-'||emp_rec.ename||'-'||emp_rec.sal);
??end loop;
?elsif choice=2 then
??dbms_output.put_line('department information');
??open_cv(2,return_cv);
??loop
??fetch return_cv into dept_rec;
??exit when return_cv%notfound;
???dbms_output.put_line(dept_rec.deptno||'-'||dept_rec.dname||'-'||dept_rec.loc);
??end loop;
?else
??raise invalid_choice;
?end if;

exception
?when invalid_choice then
??dbms_output.put_line('The CHOICE should be one of 1 and 2!');
?when others then
??dbms_output.put_line('Errors in procedure retrieve_data');
end retrieve_data;

-----------------------------------------------------------------

執(zhí)行：
SQL> exec retrieve_data(1);
employee information
7369-SMITH-800
7499-ALLEN-1600
7521-WARD-1250
7566-JONES-2975
7654-MARTIN-1250
7698-BLAKE-2850
......
PL/SQL procedure successfully completed

SQL> exec retrieve_data(2);
department information
10-ACCOUNTING-NEW YORK
20-RESEARCH-DALLAS
30-SALES-CHICAGO
40-OPERATIONS-BOSTON
PL/SQL procedure successfully completed

?

(3)用REF CURSOR實(shí)現(xiàn)BULK功能
1. To speed up INSERT, UPDATE, and DELETE statements, enclose the SQL statement within a PL/SQL FORALL statement instead of a loop construct.
2. To speed up SELECT statements, include the BULK COLLECT INTO clause in the SELECT statement instead of using INTO.

?

SQL> create table tab2 as select empno ID, ename NAME, sal SALARY from emp where 1=2;
Table created

create or replace procedure REF_BULK is
?type empcurtyp is ref cursor;
?type idlist is table of emp.empno%type;
?type namelist is table of emp.ename%type;
?type sallist is table of emp.sal%type;
?
?emp_cv empcurtyp;
?ids idlist;
?names namelist;
?sals sallist;
?
?row_cnt number;
begin
?open emp_cv for
??select empno, ename, sal from emp;
??fetch emp_cv BULK COLLECT INTO ids, names, sals; --BULK COLLECT INTO instead of INTO
?close emp_cv;

?for i in ids.first .. ids.last loop
??dbms_output.put_line('id=' || ids(i) || ' name=' || names(i) || ' salary=' || sals(i));
?end loop;

?forall i in ids.first .. ids.last --FORALL instead of FOR ...LOOP
??insert into tab2 values (ids(i), names(i), sals(i));
?commit;

?select count(*) into row_cnt from tab2;
??dbms_output.put_line('-----------------------------------');
??dbms_output.put_line('The row number of tab2 is ' || row_cnt);
?end REF_BULK;

------------------------------------------------------------

執(zhí)行：

SQL> exec ref_bulk;

id=7369 name=SMITH salary=800
id=7499 name=ALLEN salary=1600
id=7521 name=WARD salary=1250
id=7566 name=JONES salary=2975
id=7654 name=MARTIN salary=1250
id=7698 name=BLAKE salary=2850
id=7782 name=CLARK salary=2450
id=7788 name=SCOTT salary=3000
id=7839 name=KING salary=5000
id=7844 name=TURNER salary=1500
id=7876 name=ADAMS salary=1100
id=7900 name=JAMES salary=950
id=7902 name=FORD salary=3000
id=7934 name=MILLER salary=1300
-----------------------------------
The row number of tab2 is 14

PL/SQL procedure successfully completed

?

i、minus  
這個方法因?yàn)橛玫搅薽inus操作符，所以速度會受影響

select rownum,c1.* from (select t.* from carrier t where .... order by...) c1 where rownum <= 8
minus
select rownum,c2.* from (select t.* from carrier t where .... order by...) c2 where rownum <= 8

ii、select嵌套(3層)。

SELECT *
  FROM (SELECT rownum rownumber, c.*
          FROM (
                    select t.* from carrier t where .... order by .. .
                    ) c
                ) cc
 WHERE cc.rownumber > 8
   and cc.rownumber <= 10