利用DBMS_SQLTUNE优化SQL

DBMS_SQLTUNE优化SQL是在oracle10g才出来的新特性，使用它能很大程度上方便对sql的分析和优化。执行DBMS_SQLTUNE包进行sql优化需要有advisor的权限：

stat>grant advisor to stat;

Grant succeeded.

使用DBMS_SQLTUNE包进行SQL优化，大致可以分为四个步骤：

创建优化任务
执行优化任务
显示优化任务的结果
按照建议执行优化

一、创建优化任务
使用dbms_sqltune.create_tuning_task函数来创建优化任务，该参数的声明 可以通过 desc dbms_sqltune.create_tuning_task 来查看。以下是一些主要的参数的意义。
Parameter Description
sql_text The text of a SQL statement
begin_snap Begin snapshot identifier
end_snap End snapshot identifier
sql_id The identifier of a SQL statement
bind_list An ordered list of bind values in ANYDATA type
plan_hash_value The hash value of the SQL execution plan
sqlset_name The SQL tuning set name
basic_filter The SQL predicate to filter the SQL from the SQL tuning set
object_filter The object filter
rank(i) An order-by clause on the selected SQL
result_percentage A percentage on the sum of a ranking measure
result_limit The top L(imit) SQL from the (filtered/ranked) SQL
user_name The username for whom the statement is to be tuned
scope Tuning scope (limited/comprehensive)
time_limit The maximum duration in seconds for the tuning session
task_name An optional tuning task name
description A task of the SQL tuning session to a maximum of 256 characters
函数的返回值为创建的任务名。下面我们创建一个叫sql_tuning_test的任务：

代码:--------------------------------------------------------------------------------
DECLARE
my_task_name VARCHAR2 (30);
my_sqltext CLOB;
BEGIN
my_sqltext := 'select * from tj_balance_cid_200907 where clientcode=''sadfasfds''';
my_task_name := dbms_sqltune.create_tuning_task (sql_text=> my_sqltext,
bind_list => null,
user_name => 'STAT',
scope => 'COMPREHENSIVE',
time_limit => 60,
task_name => 'vega_tuning_task',
description => 'Tuning Task'
);
END;
/

PL/SQL procedure successfully completed.
--------------------------------------------------------------------------------

二、执行优化任务

通过调用dbms_sqltune.execute_tuning_task过程来执行前面创建好的优化任务。

exec dbms_sqltune.execute_tuning_task('sql_tuning_test');

PL/SQL procedure successfully completed.
--------------------------------------------------------------------------------

三、检查优化任务的状态

通过查看user_advisor_tasks/dba_advisor_tasks视图可以查看优化任务的当前状态。

代码:--------------------------------------------------------------------------------
SELECT status FROM USER_ADVISOR_TASKS WHERE task_name ='sql_tuning_test';

STATUS
-----------
COMPLETED
--------------------------------------------------------------------------------

四、得到优化任务执行的结果

通过dbms_sqltune.report_tning_task函数可以获得优化任务的结果。

代码:--------------------------------------------------------------------------------
SET LONG 999999
set serveroutput on size 999999
SET LINESIZE 1000
SELECT DBMS_SQLTUNE.REPORT_TUNING_TASK('sql_tuning_test') FROM DUAL;

DBMS_SQLTUNE.REPORT_TUNING_TAS
--------------------------------------------------------------------------------
GENERAL INFORMATION SECTION
-------------------------------------------------------------------------------
Tuning Task Name : vega_tuning_task
Tuning Task Owner : STAT
Workload Type : Single SQL Statement
Execution Count : 2
Current Execution : EXEC_4345_1
Execution Type : TUNE SQL
Scope : COMPREHENSIVE
Time Limit(seconds) : 60
Completion Status : COMPLETED
Started at : 08/04/2009 09:41:09
Completed at : 08/04/2009 09:41:10

-------------------------------------------------------------------------------
Schema Name: STAT
SQL ID : fgj1uaa7dkgt8
SQL Text : select * from tj_balance_cid_200907 where clientcode='sadfasfds'

-------------------------------------------------------------------------------
FINDINGS SECTION (1 finding)
-------------------------------------------------------------------------------

1- Index Finding (see explain plans section below)
--------------------------------------------------
通过创建一个或多个索引可以改进此语句的执行计划。

Recommendation (estimated benefit: 94.36%)
------------------------------------------
- 考虑运行可以改进物理方案设计的访问指导或者创建推荐的索引。
create index STAT.IDX$$_10F90001 on STAT.TJ_BALANCE_CID_200907("CLIENTCODE"
);

Rationale
---------
创建推荐的索引可以显著地改进此语句的执行计划。但是, 使用典型的 SQL 工作量运?
可能比单个语句更可取。通过这种方法可以获得全面的索引建议案, 包括计算索引维护

-------------------------------------------------------------------------------
EXPLAIN PLANS SECTION
-------------------------------------------------------------------------------

1- Original
-----------
Plan hash value: 494025501

--------------------------------------------------------------------------------
| Id | Operation | Name | R
--------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | |
| 1 | PARTITION LIST ALL | |
| 2 | MAT_VIEW ACCESS BY LOCAL INDEX ROWID| TJ_BALANCE_CID_200907 |
|* 3 | INDEX RANGE SCAN | IDX_TJ_BALANCE_CID_200907_01 |
--------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

3 - access("CLIENTCODE"='sadfasfds')

2- Using New Indices
--------------------
Plan hash value: 3010982149

--------------------------------------------------------------------------------
| Id | Operation | Name | Rows |
--------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 10 |
| 1 | MAT_VIEW ACCESS BY GLOBAL INDEX ROWID| TJ_BALANCE_CID_200907 | 10 |
|* 2 | INDEX RANGE SCAN | IDX$$_10F90001 | 10 |
--------------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

2 - access("CLIENTCODE"='sadfasfds')

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

然后我们就可以根据Recommendation部分的建议来执行优化操作了。

五、删除优化任务

通过调用dbms_sqltuen.drop_tuning_task可以删除已经存在的优化任务

代码:--------------------------------------------------------------------------------
exec dbms_sqltune.drop_tuning_task('sql_tuning_test');

PL/SQL procedure successfully completed.