16、数据库维护
1、管理优化程序统计信息
用于收集统计信息的首选项
DBA-> DBMS_STATS
SCOPE: statement level| table level| schema level| database level| global level
preferences: cascade | degree| estimate_percent|no_invalidate|metho_opt
granularity | incremental | publish| stale_percent
set| get| delete| export| import
exec dbms_stats.set_table_prefs(‘SH‘,‘SALES‘,‘STALE_PERCENT‘,‘13‘)
desc dbms_stats;
其中一个存储过程为：
PROCEDURE SET_TABLE_PREFS
Argument Name Type In/Out Default?
------------------------------ ----------------------- ------ --------
OWNNAME VARCHAR2 IN
TABNAME VARCHAR2 IN
PNAME VARCHAR2 IN
PVALUE VARCHAR2 IN
PROCEDURE GATHER_TABLE_STATS
Argument Name Type In/Out Default?
------------------------------ ----------------------- ------ --------
OWNNAME VARCHAR2 IN
TABNAME VARCHAR2 IN
PARTNAME VARCHAR2 IN DEFAULT
ESTIMATE_PERCENT NUMBER IN DEFAULT
BLOCK_SAMPLE BOOLEAN IN DEFAULT
METHOD_OPT VARCHAR2 IN DEFAULT
DEGREE NUMBER IN DEFAULT
GRANULARITY VARCHAR2 IN DEFAULT
CASCADE BOOLEAN IN DEFAULT
例子1：
(1)create table hr.t1 as select * from dba_objects;
因为这张表是刚刚创建的，因此看不了它的行数及block;
(2)select OWNER,TABLE_NAME,TABLESPACE_NAME,NUM_ROWS,BLOCKS from dba_tables where OWNER=‘HR‘ and TABLE_NAME=‘T1‘;
OWNER TABLE_NAME TABLESPACE NUM_ROWS BLOCKS
---------- ---------- ---------- ---------- ----------
HR T1 USERS
(3)收集统计信息
exec dbms_stats.gather_table_stats(‘HR‘,‘T1‘);
SQL> exec dbms_stats.gather_table_stats(‘HR‘,‘T1‘);
PL/SQL procedure successfully completed.
(4)继续查询，有结果
SQL> select OWNER,TABLE_NAME,TABLESPACE_NAME,NUM_ROWS,BLOCKS from dba_tables where OWNER=‘HR‘ and T
ABLE_NAME=‘T1‘;
OWNER TABLE_NAME TABLESPACE NUM_ROWS BLOCKS
HR T1 USERS ########## 1061
2、管理自动工作量资料档案库
自动工作量资料档案库（AWR）
内置性能信息资料档案库
每60分钟获取一次数据库快照，保留为8天
所有自我管理功能的基础
SGA---->MMON--->（60分钟）---》AWR
使用数据库自动诊断监视器
ADDM
在记录每个AWR快照之后运行
监事实例
在awr中存储结果。
DBMS_ADVISOR程序包
案例2：
oracle是如何给sql优化提出建议的呢？
(1)建表 create table scott.test_advisor (id varchar2(20),name varchar2(128));
(2)插入数据 insert into scott.test_advisor select object_id,object_name from dba_objects;
(3)建立索引 create index scott.idx_test_advisor_id on scott.test_advisor(id);
(4)获得统计信息
exec dbms_stats.gather_table_stats(‘SCOTT‘,‘test_advisor‘,cascade=>true);
(5)查看执行计划
set autotrace traceonly
select id,name from scott.test_advisor where id=1000;
结果为全表扫描,内容如下
SQL> select id,name from scott.test_advisor where id=1000;
Execution Plan
----------------------------------------------------------
Plan hash value: 3695065845
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time
|
| 0 | SELECT STATEMENT | | 1 | 30 | 103 (1)| 00:00:02
|
|* 1 | TABLE ACCESS FULL| TEST_ADVISOR | 1 | 30 | 103 (1)| 00:00:02
|
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter(TO_NUMBER("ID")=1000)
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
375 consistent gets
0 physical reads
0 redo size
596 bytes sent via SQL*Net to client
524 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
下面定义任务，看看oracle给sql有什么的好建议呢？
(6)定义任务
SQL> declare
2 l_task_id varchar2(20);
3 l_sql varchar2(2000);
4 begin
5 l_sql :=‘select id,name from scott.test_advisor where id=1000‘;
6 l_task_id :=dbms_sqltune.create_tuning_task(
7 sql_text =>l_sql,
8 user_name =>‘SCOTT‘,
9 scope =>‘COMPREHENSIVE‘,
10 time_limit =>30,
11 task_name =>‘manual_advisor‘);
12 end;
13 /
PL/SQL procedure successfully completed.
(7)执行任务
SQL> begin
2 dbms_sqltune.execute_tuning_task(‘manual_advisor‘);
3 end;
4 /
PL/SQL procedure successfully completed.
(8)查看建议
set serveroutput on size 99999;
set long 99999;
select dbms_sqltune.report_tuning_task(‘manual_advisor‘) from dual;
建议结果如下：
SQL> select dbms_sqltune.report_tuning_task(‘manual_advisor‘) from dual;
DBMS_SQLTUNE.REPORT_TUNING_TASK(‘MANUAL_ADVISOR‘)
--------------------------------------------------------------------------------
GENERAL INFORMATION SECTION
-------------------------------------------------------------------------------
Tuning Task Name : manual_advisor
Tuning Task Owner : SYS
Workload Type : Single SQL Statement
Scope : COMPREHENSIVE
Time Limit(seconds): 30
Completion Status : COMPLETED
Started at : 07/02/2017 21:21:56
Completed at : 07/02/2017 21:21:59
DBMS_SQLTUNE.REPORT_TUNING_TASK(‘MANUAL_ADVISOR‘)
Schema Name: SCOTT
SQL ID : 7gns85v297ncy
SQL Text : select id,name from scott.test_advisor where id=1000
FINDINGS SECTION (2 findings)
1- Index Finding (see explain plans section below)
DBMS_SQLTUNE.REPORT_TUNING_TASK(‘MANUAL_ADVISOR‘)
The execution plan of this statement can be improved by creating one or more
indices.
Recommendation (estimated benefit: 98.04%)
------------------------------------------
- Consider running the Access Advisor to improve the physical schema design
or creating the recommended index.
create index SCOTT.IDX$$_023A0001 on SCOTT.TEST_ADVISOR(TO_NUMBER("ID"));
Rationale
---------
DBMS_SQLTUNE.REPORT_TUNING_TASK(‘MANUAL_ADVISOR‘)
--------------------------------------------------------------------------------
Creating the recommended indices significantly improves the execution plan
of this statement. However, it might be preferable to run "Access Advisor"
using a representative SQL workload as opposed to a single statement. This
will allow to get comprehensive index recommendations which takes into
account index maintenance overhead and additional space consumption.
2- Restructure SQL finding (see plan 1 in explain plans section)
----------------------------------------------------------------
The predicate TO_NUMBER("TEST_ADVISOR"."ID")=1000 used at line ID 1 of the
execution plan contains an implicit data type conversion on indexed column
"ID". This implicit data type conversion prevents the optimizer from
从上面可以看到有2条建议，说可以提高性能的98.4%(Recommendation (estimated benefit: 98.04%))分别如下：
create index SCOTT.IDX$$_023A0001 on SCOTT.TEST_ADVISOR(TO_NUMBER("ID"));
TO_NUMBER("TEST_ADVISOR"."ID")=1000
下面来看看第二条建议怎么样？
select id,name from scott.test_advisor where TO_NUMBER("TEST_ADVISOR"."ID")=1000;
结果还是全表扫描cost为103
在看看第一条建议
create index SCOTT.IDX$$_023A0001 on SCOTT.TEST_ADVISOR(TO_NUMBER("ID"));
在查询结果为走索引，代价为3，的确提高了98.4%性能！！！！！！！