SQL Tuning Advisor使用实例

在oracle10g之前，想要优化一个sql语句是比较麻烦，但是在oracle10g这个版本推出的SQL Tuning Advisor这个工具，能大大减少sql调优的工作量，不过要想使用SQL Tuning Advisor，一定要保证你的优化器是CBO模式。

1.首先需要创建一个用于调优的用户bamboo,并授予advisor给创建的用户
SQL> create user bamboo identified by bamboo;
User created.

SQL> grant connect,resource to bamboo;
Grant succeeded.

SQL> grant advisor to bamboo;
Grant succeeded.

2.创建用户做测试的2张表,大表里面插入500万条数据，小表里面插入10万条数据，其创建方法如下
SQL> create table bigtable (id number(10),name varchar2(100));
Table created.

SQL> begin
  2  for i in 1..5000000 loop
  3  insert into bigtable values(i,'test'||i);
  4  end loop;
  5  end;
  6  /

PL/SQL procedure successfully completed.

SQL> commit;

SQL> create table smalltable (id number(10),name varchar2(100));
Table created.

SQL> begin
  2  for i in 1..100000 loop
  3  insert into smalltable values(i,'test'||i);
  4  end loop;
  5  end;
  6  /

PL/SQL procedure successfully completed.

SQL> commit;

3.然后对bigtable和smalltable做一个等连接查询，然后跟踪其执行计划
SQL> select a.id,a.name,b.id,b.name from bigtable a,smalltable b where a.id=b.id and a.id=40000;

        ID        NAME               ID         NAME
---------- -------------------- ---------- ---------------------------
     40000      test40000          40000      test40000

Execution Plan
----------------------------------------------------------
Plan hash value: 1703851322

---------------------------------------------------------------------------------
| Id  | Operation          | Name       | Rows  | Bytes | Cost (%CPU)| Time     |
---------------------------------------------------------------------------------
|   0 | SELECT STATEMENT   |            |   839 |   106K|  3656   (5)| 00:00:44 |
|*  1 |  HASH JOIN         |            |   839 |   106K|  3656   (5)| 00:00:44 |
|*  2 |   TABLE ACCESS FULL| SMALLTABLE |     5 |   325 |    71   (3)| 00:00:01 |
|*  3 |   TABLE ACCESS FULL| BIGTABLE   |   173 | 11245 |  3584   (5)| 00:00:44 |
---------------------------------------------------------------------------------

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

   1 - access("A"."ID"="B"."ID")
   2 - filter("B"."ID"=40000)
   3 - filter("A"."ID"=40000)

Note
-----
   - dynamic sampling used for this statement

Statistics
----------------------------------------------------------
          9  recursive calls
          0  db block gets
      16151  consistent gets
      11469  physical reads
          0  redo size
        588  bytes sent via SQL*Net to client
        385  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          2  sorts (memory)
          0  sorts (disk)
          1  rows processed

熟悉执行计划的就可以看出，这个sql执行是很慢的，2个表都做的是全表扫描，并且其物理读是11469，按照优化的经验，给2个表的id创建索引，减少查询时候的物理读,下面我们就看看通过优化器，oracle能我们什么样的建议呢？

4.下面就通过DBMS_SQLTUNE包的CREATE_TUNING_TASK来创建一个优化任务，然后通过DBMS_SQLTUNE.EXECUTE_TUNING_TASK来执行调优任务，生成调优建议

SQL> DECLARE 
  2    my_task_name VARCHAR2(30); 
  3    my_sqltext CLOB; 
  4  BEGIN 
  5    my_sqltext := 'select a.id,a.name,b.id,b.name from bigtable a,smalltable b where a.id=b.id and a.id=40000'; 
  6  
  7    my_task_name := DBMS_SQLTUNE.CREATE_TUNING_TASK( 
  8                            sql_text => my_sqltext, 
  9                            user_name => 'SCOTT', 
10                             scope => 'COMPREHENSIVE', 
11                             time_limit => 60, 
12                             task_name => 'test_sql_tuning_task1', 
13                             description => 'Task to tune a query'); 
14     DBMS_SQLTUNE.EXECUTE_TUNING_TASK(task_name => 'test_sql_tuning_task1');
15  END; 
16  /

5.执行的过程中，也可以通过user_advisor_tasks或者dba_advisor_tasks来查看调优任务执行的状况

SQL> select task_name,ADVISOR_NAME,STATUS from user_advisor_tasks;

TASK_NAME                      ADVISOR_NAME                             STATUS
------------------ ---------------------------------------- ---------------------------------
test_sql_tuning_task1          SQL Tuning Advisor                       COMPLETED

如果status是EXECUTING，则表示任务正在执行，如果为COMPLETED，则任务已经执行完毕

6.通过调用dbms_sqltune.report_tuning_task可以查询调优的结果，不过在查询结果之前，得设置sqlplus的环境，如果不设置，则查询的结果出不来
SQL> set long 999999
SQL> set LONGCHUNKSIZE 999999
SQL> set serveroutput on size 999999
SQL> set linesize 200
SQL> select dbms_sqltune.report_tuning_task('test_sql_tuning_task1') from dual;

SQL> select dbms_sqltune.report_tuning_task('test_sql_tuning_task1') from dual;

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
---------------------------------------------------------------------------------------------------------------------------------
GENERAL INFORMATION SECTION
-------------------------------------------------------------------------------
Tuning Task Name                  : test_sql_tuning_task1
Tuning Task Owner                 : BAMBOO
Scope                             : COMPREHENSIVE
Time Limit(seconds)               : 60
Completion Status                 : COMPLETED
Started at                        : 10/13/2011 05:07:53
Completed at                      : 10/13/2011 05:08:18
Number of Statistic Findings      : 2
Number of Index Findings          : 1

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
----------------------------------------------------------------------------------------------------------------------------------
Schema Name: SCOTT
SQL ID     : 7arau1k5a3mv1
SQL Text   : select a.id,a.name,b.id,b.name from bigtable a,smalltable b
             where a.id=b.id and a.id=40000

-------------------------------------------------------------------------------
FINDINGS SECTION (3 findings)
-------------------------------------------------------------------------------

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
----------------------------------------------------------------------------------------------------------------------------------
1- Statistics Finding
---------------------
  Table "SCOTT"."SMALLTABLE" was not analyzed.

  Recommendation
  --------------
  - Consider collecting optimizer statistics for this table.
    execute dbms_stats.gather_table_stats(ownname => 'SCOTT', tabname =>
            'SMALLTABLE', estimate_percent => DBMS_STATS.AUTO_SAMPLE_SIZE,
            method_opt => 'FOR ALL COLUMNS SIZE AUTO');

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
----------------------------------------------------------------------------------------------------------------------------------
  Rationale
  ---------
    The optimizer requires up-to-date statistics for the table in order to
    select a good execution plan.

2- Statistics Finding
---------------------
  Table "SCOTT"."BIGTABLE" was not analyzed.

  Recommendation
  --------------

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
----------------------------------------------------------------------------------------------------------------------------------
  - Consider collecting optimizer statistics for this table.
    execute dbms_stats.gather_table_stats(ownname => 'SCOTT', tabname =>
            'BIGTABLE', estimate_percent => DBMS_STATS.AUTO_SAMPLE_SIZE,
            method_opt => 'FOR ALL COLUMNS SIZE AUTO');

  Rationale
  ---------
    The optimizer requires up-to-date statistics for the table in order to
    select a good execution plan.

3- Index Finding (see explain plans section below)

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
---------------------------------------------------------------------------------------------------------------------------------
  The execution plan of this statement can be improved by creating one or more
  indices.

  Recommendation (estimated benefit: 100%)
  ----------------------------------------
  - Consider running the Access Advisor to improve the physical schema design
    or creating the recommended index.
    create index SCOTT.IDX$$_00790001 on SCOTT.SMALLTABLE('ID');

  - Consider running the Access Advisor to improve the physical schema design

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
----------------------------------------------------------------------------------------------------------------------------------
    or creating the recommended index.
    create index SCOTT.IDX$$_00790002 on SCOTT.BIGTABLE('ID');

  Rationale
  ---------
    Creating the recommended indices significantly improves the execution plan
    of this statement. However, it might be preferable to run "Access Advisor"
    using a representative SQL workload as opposed to a single statement. This
    will allow to get comprehensive index recommendations which takes into
    account index maintenance overhead and additional space consumption.

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
----------------------------------------------------------------------------------------------------------------------------------
-------------------------------------------------------------------------------
EXPLAIN PLANS SECTION
-------------------------------------------------------------------------------

1- Original
-----------
Plan hash value: 1703851322

---------------------------------------------------------------------------------
| Id  | Operation          | Name       | Rows  | Bytes | Cost (%CPU)| Time     |
---------------------------------------------------------------------------------

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
----------------------------------------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT   |            |   839 |   106K|  3656   (5)| 00:00:44 |
|*  1 |  HASH JOIN         |            |   839 |   106K|  3656   (5)| 00:00:44 |
|*  2 |   TABLE ACCESS FULL| SMALLTABLE |     5 |   325 |    71   (3)| 00:00:01 |
|*  3 |   TABLE ACCESS FULL| BIGTABLE   |   173 | 11245 |  3584   (5)| 00:00:44 |
---------------------------------------------------------------------------------

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

   1 - access("A"."ID"="B"."ID")
   2 - filter("B"."ID"=40000)

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
---------------------------------------------------------------------------------------------------------------------------------
   3 - filter("A"."ID"=40000)

2- Using New Indices
--------------------
Plan hash value: 3720188830

------------------------------------------------------------------------------------------------
| Id  | Operation                     | Name           | Rows  | Bytes | Cost (%CPU)| Time     |
------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT              |                |     1 |   130 |     5   (0)| 00:00:01 |
|   1 |  TABLE ACCESS BY INDEX ROWID  | BIGTABLE       |     1 |    65 |     3   (0)| 00:00:01 |

DBMS_SQLTUNE.REPORT_TUNING_TASK('TEST_SQL_TUNING_TASK1')
---------------------------------------------------------------------------------------------------------------------------------
|   2 |   NESTED LOOPS                |                |     1 |   130 |     5   (0)| 00:00:01 |
|   3 |    TABLE ACCESS BY INDEX ROWID| SMALLTABLE     |     1 |    65 |     2   (0)| 00:00:01 |
|*  4 |     INDEX RANGE SCAN          | IDX$$_00790001 |     1 |       |     1   (0)| 00:00:01 |
|*  5 |    INDEX RANGE SCAN           | IDX$$_00790002 |     1 |       |     2   (0)| 00:00:01 |
------------------------------------------------------------------------------------------------

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

   4 - access("B"."ID"=40000)
   5 - access("A"."ID"=40000)

  从上面的结果可以看到oracle的调优顾问给我们3条建议：
（1）SCOTT.SMALLTABLE表没有做分析，需要做一下表结构的分析，并且给出一个分析的建议，如下所示
     execute dbms_stats.gather_table_stats(ownname => 'SCOTT', tabname =>
            'SMALLTABLE', estimate_percent => DBMS_STATS.AUTO_SAMPLE_SIZE,
            method_opt => 'FOR ALL COLUMNS SIZE AUTO');
（2）SCOTT.BIGTABLE表没有做分析，需要做一下表结构的分析，并且给出一个分析的建议，如下所示
     execute dbms_stats.gather_table_stats(ownname => 'SCOTT', tabname =>
            'BIGTABLE', estimate_percent => DBMS_STATS.AUTO_SAMPLE_SIZE,
            method_opt => 'FOR ALL COLUMNS SIZE AUTO');
（3）oracle建议我们在表SCOTT.SMALLTABLE，SCOTT.BIGTABLE的id列创建一个bitree索引，给的建议如下
      create index SCOTT.IDX$$_00790002 on SCOTT.BIGTABLE('ID');  
      create index SCOTT.IDX$$_00790001 on SCOTT.SMALLTABLE('ID');
    当然创建索引的名字可以改成别的名字
   

通过以上查看oracle的调优顾问给的建议，基本和我们在前面给出的调优方案是一致，因此当我们给一个大的ＳＱＬ做优化的时候，可以先使用oracle调优顾问，得到一些调优方案，然后根据实际情况做一些调整就可以。

 以下就是执行oracle调优顾问的建议，重新执行select a.id,a.name,b.id,b.name from bigtable a,smalltable b where a.id=b.id and a.id=40000这条语句得到的执行计划，可以看出查询时间和物理读大大减少
 SQL> select a.id,a.name,b.id,b.name from bigtable a,smalltable b where a.id=b.id and a.id=40000;

        ID   NAME               ID     NAME
---------- ---------     ---------- ----------------------------------------
     40000 test40000           40000   test40000

Execution Plan
----------------------------------------------------------
Plan hash value: 777647921

-------------------------------------------------------------------------------------------------
| Id  | Operation                     | Name            | Rows  | Bytes | Cost (%CPU)| Time     |
-------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT              |                 |     1 |    31 |     5   (0)| 00:00:01 |
|   1 |  TABLE ACCESS BY INDEX ROWID  | BIGTABLE        |     1 |    17 |     3   (0)| 00:00:01 |
|   2 |   NESTED LOOPS                |                 |     1 |    31 |     5   (0)| 00:00:01 |
|   3 |    TABLE ACCESS BY INDEX ROWID| SMALLTABLE      |     1 |    14 |     2   (0)| 00:00:01 |
|*  4 |     INDEX RANGE SCAN          | I_ID_SAMLLTABLE |     1 |       |     1   (0)| 00:00:01 |
|*  5 |    INDEX RANGE SCAN           | I_ID_BIGTABLE   |     1 |       |     2   (0)| 00:00:01 |
-------------------------------------------------------------------------------------------------

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

   4 - access("B"."ID"=40000)
   5 - access("A"."ID"=40000)

Statistics
----------------------------------------------------------
          0  recursive calls
          0  db block gets
          9  consistent gets
          0  physical reads
          0  redo size
        588  bytes sent via SQL*Net to client
        385  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
          1  rows processed

转自：http://www.cnblogs.com/einyboy/archive/2012/08/14/2637659.html