dbms_stats包dbms_stats.gather_table_stats&动态采样

最新推荐文章于 2024-05-06 17:55:27 发布

zml19910422

最新推荐文章于 2024-05-06 17:55:27 发布

阅读量825

点赞数

分类专栏： sql/plsql优化文章标签： dbms_stats gather_table_stats 动态采样 dynamic sampling

本文链接：https://blog.csdn.net/zml19910422/article/details/23863791

版权

sql/plsql优化专栏收录该内容

9 篇文章 0 订阅

订阅专栏

dbms_stats包中的主要存储过程：参考博客

1、dbms_stats.gather_table_stats

----DBMS_STATS.GATHER_TABLE_STATS实例
SQL> create table t as select * from dba_objects;

表已创建。

SQL> create index ind_t_type on t(object_type);

索引已创建。

SQL> select num_rows from dba_tables where table_name = 'T';

  NUM_ROWS
----------


SQL> set autot trace exp
SQL> select object_id,object_name from t where object_type = 'TABLE';

执行计划
----------------------------------------------------------
Plan hash value: 3351898915

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

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

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

|   0 | SELECT STATEMENT            |            |  2500 |   219K|   110   (0)|00:00:02 |

|   1 |  TABLE ACCESS BY INDEX ROWID| T          |  2500 |   219K|   110   (0)|00:00:02 |

|*  2 |   INDEX RANGE SCAN          | IND_T_TYPE |  2500 |       |     7   (0)|00:00:01 |

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


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

   2 - access("OBJECT_TYPE"='TABLE')

Note
-----
   - dynamic sampling used for this statement (level=2)--因表T没有被分析，所以此处用了动态采样统计信息，因用了动态采样，使rows值接近于真实值。

SQL> set autot off
SQL> select num_rows from dba_tables where table_name = 'T';

  NUM_ROWS
----------


SQL> exec dbms_stats.gather_table_stats('SYSTEM','T',estimate_percent => 100,cascade => true,method_opt => 'FOR ALL COLUMNS SIZE AUTO');

PL/SQL 过程已成功完成。

SQL> select num_rows from dba_tables where table_name = 'T';

  NUM_ROWS
----------
     72473

SQL> SET AUTOT TRACE EXP;
SQL> select object_id,object_name from t where object_type = 'TABLE';

执行计划
----------------------------------------------------------
Plan hash value: 3351898915

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

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

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

|   0 | SELECT STATEMENT            |            |  2856 |   108K|   127   (0)|00:00:02 |

|   1 |  TABLE ACCESS BY INDEX ROWID| T          |  2856 |   108K|   127   (0)|00:00:02 |

|*  2 |   INDEX RANGE SCAN          | IND_T_TYPE |  2856 |       |     8   (0)|00:00:01 |

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


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

   2 - access("OBJECT_TYPE"='TABLE')

SQL> select count(*) from t where object_type = 'TABLE';  
COUNT(*)
----------
      2856
SQL>

使用gather_table_stats前dba_tables中的num_rows字段为空，收集表的统计信息之后num_rows的值为表的行数。

执行计划中rows值在收集统计信息之后较为接近真实行数，使优化器在选择更优的执行计划

收集统计信息使优化器选择更优的执行计划。

2、动态采样dynamic sampling 参考博客

---动态采样实例，/*+ dynamic_sampling(h 0) */  动态采样级别设置为0表示不做动态采样
SQL> create table h as select * from dba_tables;

表已创建。

SQL> select /*+ dynamic_sampling(h 0) */ * from h;----不做动态采样

执行计划
----------------------------------------------------------
Plan hash value: 3725287857

--------------------------------------------------------------------------
| Id  | Operation         | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |      |  7923 |  3992K|    29   (4)| 00:00:01 |
|   1 |  TABLE ACCESS FULL| H    |  7923 |  3992K|    29   (4)| 00:00:01 |
--------------------------------------------------------------------------

SQL> select count(*) from h;----动态采样

执行计划
----------------------------------------------------------
Plan hash value: 3272090800

-------------------------------------------------------------------
| Id  | Operation          | Name | Rows  | Cost (%CPU)| Time     |
-------------------------------------------------------------------
|   0 | SELECT STATEMENT   |      |     1 |    28   (0)| 00:00:01 |
|   1 |  SORT AGGREGATE    |      |     1 |            |          |
|   2 |   TABLE ACCESS FULL| H    |  2771 |    28   (0)| 00:00:01 |
-------------------------------------------------------------------

Note
-----
   - dynamic sampling used for this statement (level=2)----动态采样

SQL>SQL> set autot off
SQL> select count(*) from h;
COUNT(*)
----------
      2783
SQL>

可以看出使用动态采样rows的值更接近于真实值，当表没有被分析时使用动态采样。