ORACLE绑定变量BIND PEEKING
ORACLE 在9i之后引入了bind peeking，通过bind peeking，oracle可以在硬解析的时候窥探绑定变量的值，并根据当前绑定变量的值生成执行计划。在oracle 9i之前的版本中，oracle仅仅通过统计信息来生成执行计划。
下面看一下不同版本oracle下绑定变量对执行计划的影响
SQL> alter system flush shared_pool;系统已更改。SQL> alter system set optimizer_features_enable='8.1.7';系统已更改。SQL> var v number;
SQL> exec :v := 1;PL/SQL 过程已成功完成。SQL> select count(*) from acs_test_tab where record_type = :v;COUNT(*)
----------1SQL> select * from table(dbms_xplan.display_cursor(null,null,'advanced'));PLAN_TABLE_OUTPUT
----------------------------------------------------------------------------------------------------
SQL_ID 3rg5r8sghcvb3, child number 0
-------------------------------------
select count(*) from acs_test_tab where record_type = :vPlan hash value: 2956728990--------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost |
--------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | 3 |
| 1 | SORT AGGREGATE | | 1 | 4 | |
|* 2 | INDEX RANGE SCAN| ACS_TEST_TAB_RECORD_TYPE_I | 2 | 8 | 3 |
--------------------------------------------------------------------------------Predicate Information (identified by operation id):
---------------------------------------------------2 - access("RECORD_TYPE"=:V)已选择47行。SQL> alter system flush shared_pool;系统已更改。SQL> alter system set optimizer_features_enable='11.2.0.3.1';系统已更改。SQL> var v number;
SQL> exec :v := 1;PL/SQL 过程已成功完成。SQL> select count(*) from acs_test_tab where record_type = :v;COUNT(*)
----------1SQL> select * from table(dbms_xplan.display_cursor(null,null,'advanced'));PLAN_TABLE_OUTPUT
----------------------------------------------------------------------------------------------------
SQL_ID 3rg5r8sghcvb3, child number 0
-------------------------------------
select count(*) from acs_test_tab where record_type = :vPlan hash value: 2956728990------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | 3 (100)| |
| 1 | SORT AGGREGATE | | 1 | 4 | | |
|* 2 | INDEX RANGE SCAN| ACS_TEST_TAB_RECORD_TYPE_I | 1 | 4 | 3 (0)| 00:00:01 |
------------------------------------------------------------------------------------------------Peeked Binds (identified by position):
--------------------------------------1 - :V (NUMBER): 1 --绑定变量窥探Predicate Information (identified by operation id):
---------------------------------------------------2 - access("RECORD_TYPE"=:V)已选择49行。SQL> alter system flush shared_pool;系统已更改。SQL> exec :v := 2;PL/SQL 过程已成功完成。SQL> select count(*) from acs_test_tab where record_type = :v;COUNT(*)
----------50000SQL> select * from table(dbms_xplan.display_cursor(null,null,'advanced'));PLAN_TABLE_OUTPUT
----------------------------------------------------------------------------------------------------
SQL_ID 3rg5r8sghcvb3, child number 0
-------------------------------------
select count(*) from acs_test_tab where record_type = :vPlan hash value: 2957754476----------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | 136 (100)| |
| 1 | SORT AGGREGATE | | 1 | 4 | | |
|* 2 | INDEX FAST FULL SCAN| ACS_TEST_TAB_RECORD_TYPE_I | 48031 | 187K| 136 (1)| 00:00:02 |
----------------------------------------------------------------------------------------------------Peeked Binds (identified by position):
--------------------------------------1 - :V (NUMBER): 2 --绑定变量窥探，绑定变量会影响最初硬解析的执行计划Predicate Information (identified by operation id):
---------------------------------------------------2 - filter("RECORD_TYPE"=:V)已选择49行。使用绑定变量窥测的好处是：可以帮助优化器在第一次硬解析时选择最优的执行计划。但是同时这也是其弊端：在第一次硬解析后，后面发生的所有解析都会使用第一次硬解析生成的执行计划，如果数据的分布是均匀的，问题不大，如果数据分布式倾斜的，那么第一次硬解析生成的执行计划未必是最优的，甚至可能是非常糟糕的。例如：
SQL> show parameter optimizer_featNAME TYPE VALUE
------------------------------------ ----------- ------------------------------
optimizer_features_enable string 11.2.0.3.1
SQL> alter system flush shared_pool;系统已更改。SQL> var v number;
SQL> exec :v := 2;PL/SQL 过程已成功完成。SQL> select count(*) from acs_test_tab where record_type = :v;COUNT(*)
----------50000SQL> select * from table(dbms_xplan.display_cursor);PLAN_TABLE_OUTPUT
----------------------------------------------------------------------------------------------------
SQL_ID 3rg5r8sghcvb3, child number 0
-------------------------------------
select count(*) from acs_test_tab where record_type = :vPlan hash value: 2957754476----------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | 136 (100)| |
| 1 | SORT AGGREGATE | | 1 | 4 | | |
|* 2 | INDEX FAST FULL SCAN| ACS_TEST_TAB_RECORD_TYPE_I | 48031 | 187K| 136 (1)| 00:00:02 |
----------------------------------------------------------------------------------------------------Predicate Information (identified by operation id):
---------------------------------------------------2 - filter("RECORD_TYPE"=:V)已选择19行。SQL> exec :v := 1PL/SQL 过程已成功完成。SQL> select count(*) from acs_test_tab where record_type = :v;COUNT(*)
----------1SQL> select * from table(dbms_xplan.display_cursor);PLAN_TABLE_OUTPUT
----------------------------------------------------------------------------------------------------
SQL_ID 3rg5r8sghcvb3, child number 0
-------------------------------------
select count(*) from acs_test_tab where record_type = :vPlan hash value: 2957754476----------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | 136 (100)| |
| 1 | SORT AGGREGATE | | 1 | 4 | | |
|* 2 | INDEX FAST FULL SCAN| ACS_TEST_TAB_RECORD_TYPE_I | 48031 | 187K| 136 (1)| 00:00:02 |
----------------------------------------------------------------------------------------------------Predicate Information (identified by operation id):
---------------------------------------------------2 - filter("RECORD_TYPE"=:V)已选择19行。SQL> select count(*) from acs_test_tab where record_type = 1;COUNT(*)
----------1SQL> select * from table(dbms_xplan.display_cursor);PLAN_TABLE_OUTPUT
----------------------------------------------------------------------------------------------------
SQL_ID 1pxm87f6yd0bp, child number 0
-------------------------------------
select count(*) from acs_test_tab where record_type = 1Plan hash value: 2956728990------------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
------------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | 3 (100)| |
| 1 | SORT AGGREGATE | | 1 | 4 | | |
|* 2 | INDEX RANGE SCAN| ACS_TEST_TAB_RECORD_TYPE_I | 1 | 4 | 3 (0)| 00:00:01 |
------------------------------------------------------------------------------------------------Predicate Information (identified by operation id):
---------------------------------------------------2 - access("RECORD_TYPE"=1)已选择19行。对于变量v的取值为1的执行计划和采用常量1的执行计划性能差距还是比较大的。
总结：oracle在9i后引入变量窥测技术，该技术对于数据分布均匀的数据是非常合适的，但是对于分布倾斜的数据或者在OLAP系统中是不建议使用的。
本文来自互联网用户投稿，文章观点仅代表作者本人，不代表本站立场，不承担相关法律责任。如若转载，请注明出处。 如若内容造成侵权/违法违规/事实不符，请点击【内容举报】进行投诉反馈！