ORACLE WHER条件的执行顺序是不是自右向左

突然在网上看到一种说法，ORACLE的WHERE条件执行顺序是自右向左的。

理由是，当ORACLE的WHERE条件中出现多个00904表示符无效错误时，错误是从右向左的顺序报的。

也有人提出解析顺序和执行顺序不是一码事，执行顺序要看执行计划。

于是，我特地作了个试验。

现做一个试验用表

SQL> select count(*) from dba_objects;


  COUNT(*)
----------
     72536

create table dba_objects2 as select * from dba_objects;

先来一次查询

SQL> select object_id,object_name from dba_objects2 where owner='HR' and object_id>500;


已选择41行。




执行计划
----------------------------------------------------------
Plan hash value: 4278809457


----------------------------------------------------------------------------------
| Id  | Operation         | Name         | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |              |   426 | 40896 |   283   (1)| 00:00:04 |
|*  1 |  TABLE ACCESS FULL| DBA_OBJECTS2 |   426 | 40896 |   283   (1)| 00:00:04 |
----------------------------------------------------------------------------------


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


   1 - filter("OWNER"='HR' AND "OBJECT_ID">500)


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

可以看出，这次Oracle是自左向右执行。那么SQL语句颠倒下会怎样？

SQL> select object_id,object_name from dba_objects2 where object_id>500 and owner='HR';


已选择41行。




执行计划
----------------------------------------------------------
Plan hash value: 4278809457


----------------------------------------------------------------------------------
| Id  | Operation         | Name         | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |              |   426 | 40896 |   283   (1)| 00:00:04 |
|*  1 |  TABLE ACCESS FULL| DBA_OBJECTS2 |   426 | 40896 |   283   (1)| 00:00:04 |
----------------------------------------------------------------------------------


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


   1 - filter("OWNER"='HR' AND "OBJECT_ID">500)


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

实际的执行顺序是自右向左。难道oracle的执行顺序是看开销的？把尽可能多缩小数据范围的条件放在前面执行。

为了证明自己的想法，我又增加了一个条件。

SQL> select object_id,object_name from dba_objects2 where object_id>500 and object_name like '%EM%'  and owner='HR';


已选择14行。




执行计划
----------------------------------------------------------
Plan hash value: 4278809457


----------------------------------------------------------------------------------
| Id  | Operation         | Name         | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |              |   147 | 14112 |   283   (1)| 00:00:04 |
|*  1 |  TABLE ACCESS FULL| DBA_OBJECTS2 |   147 | 14112 |   283   (1)| 00:00:04 |
----------------------------------------------------------------------------------


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


   1 - filter("OBJECT_NAME" IS NOT NULL AND "OWNER"='HR' AND
              "OBJECT_ID">500 AND "OBJECT_NAME" LIKE '%EM%')


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

ORACLE 智能的增加了"OBJECT_NAME" IS NOT NULL的条件，把放在中间的 OBJECT_NAME" LIKE '%EM%'放在了最后执行。

由此可见，WHERE条件的执行顺序是按照开销的右小到大的顺序，智能调节的。