oracle hint_noappend_parallel

1,hint
model_min_analysis
 1,/*+ model_min_analysis */
 2,忽略一些spreadsheet rules的编译期间的优化工作,这个主要指的具体有依赖性关系的图表分析
 3,其它一些spreadsheet优化,比如构建filter产生spreadsheet访问结构和有限的规则pruning
 
 
 noappend
 1,让插入表时通过非并行模式且以传统方式插入数据
 3,传统方式是串行方式插入数据
 2,direct path insert是并行模式插入
 4,何谓direct path insert
 
 SQL> create table t_append(a int);

表已创建。

SQL> explain plan for insert into t_append select level from dual connect by lev
el<=1e5;

已解释。

SQL> select * from table(dbms_xplan.display);

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------

Plan hash value: 1236776825

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

--

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

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

--

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------

|   0 | INSERT STATEMENT              |          |     1 |     2   (0)| 00:00:01

 |

|   1 |  LOAD TABLE CONVENTIONAL      | T_APPEND |       |            |
 |

|*  2 |   CONNECT BY WITHOUT FILTERING|          |       |            |
 |

|   3 |    FAST DUAL                  |          |     1 |     2   (0)| 00:00:01

 |

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------

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

--

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

   2 - filter(LEVEL<=1e5)

已选择15行。

SQL> explain plan for insert /*+ append */ into t_append select level from dual
connect by level<=1e5;

已解释。

SQL> select * from table(dbms_xplan.display);

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------

Plan hash value: 1541388231

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

--

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

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

--

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------

|   0 | INSERT STATEMENT              |          |     1 |     2   (0)| 00:00:01

 |

|   1 |  LOAD AS SELECT               | T_APPEND |       |            |
 |

|*  2 |   CONNECT BY WITHOUT FILTERING|          |       |            |
 |

|   3 |    FAST DUAL                  |          |     1 |     2   (0)| 00:00:01

 |

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------

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

--

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

   2 - filter(LEVEL<=1e5)

已选择15行。

来自 “ ITPUB博客 ” ，链接：http://blog.itpub.net/9240380/viewspace-751382/，如需转载，请注明出处，否则将追究法律责任。

转载于:http://blog.itpub.net/9240380/viewspace-751382/