其实RULE优化模式已经基本上退出了历史舞台，不过既然有人提起了，就把概念描述清楚，对RULE优化模式不感兴趣的可以无视这篇文章。

RULE方式是基于规则的，根据FROM列表中的顺序判断驱动表是规则之一，但是这个规则的优先级并不高，如果有其他优先级规则存在，则执行计划会发生改变：

SQL> CREATE TABLE T1            
  2  (ID NUMBER,
  3  NAME VARCHAR2(30));

Table created.

SQL> CREATE TABLE T2
  2  (ID NUMBER,
  3  NAME VARCHAR2(30));

Table created.

SQL> INSERT INTO T1
  2  SELECT ROWNUM, TNAME     
  3  FROM TAB;

95 rows created.

SQL> INSERT INTO T2
  2  SELECT ROWNUM, TABLE_NAME
  3  FROM DBA_TABLES;

8923 rows created.

SQL> COMMIT;

Commit complete.

下面将优化模式改为RULE模式，看看是否驱动表的选择与FROM列表中的位置有关：

SQL> ALTER SESSION SET OPTIMIZER_MODE = RULE;

Session altered.

SQL> SET AUTOT TRACE
SQL> SELECT *
  2  FROM T1, T2
  3  WHERE T1.ID = T2.ID;

95 rows selected.

Execution Plan
----------------------------------------------------------
Plan hash value: 1792967693

------------------------------------
| Id  | Operation           | Name |
------------------------------------
|   0 | SELECT STATEMENT    |      |
|   1 |  MERGE JOIN         |      |
|   2 |   SORT JOIN         |      |
|   3 |    TABLE ACCESS FULL| T2   |
|*  4 |   SORT JOIN         |      |
|   5 |    TABLE ACCESS FULL| T1   |
------------------------------------

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

   4 - access("T1"."ID"="T2"."ID")
       filter("T1"."ID"="T2"."ID")

Note
-----
   - rule based optimizer used (consider using cbo)

Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
         30  consistent gets
          0  physical reads
          0  redo size
       4545  bytes sent via SQL*Net to client
        558  bytes received via SQL*Net from client
          8  SQL*Net roundtrips to/from client
          2  sorts (memory)
          0  sorts (disk)
         95  rows processed

SQL> SELECT *
  2  FROM T2, T1
  3  WHERE T1.ID = T2.ID;

95 rows selected.

Execution Plan
----------------------------------------------------------
Plan hash value: 412793182

------------------------------------
| Id  | Operation           | Name |
------------------------------------
|   0 | SELECT STATEMENT    |      |
|   1 |  MERGE JOIN         |      |
|   2 |   SORT JOIN         |      |
|   3 |    TABLE ACCESS FULL| T1   |
|*  4 |   SORT JOIN         |      |
|   5 |    TABLE ACCESS FULL| T2   |
------------------------------------

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

   4 - access("T1"."ID"="T2"."ID")
       filter("T1"."ID"="T2"."ID")

Note
-----
   - rule based optimizer used (consider using cbo)

Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
         30  consistent gets
          0  physical reads
          0  redo size
       4545  bytes sent via SQL*Net to client
        558  bytes received via SQL*Net from client
          8  SQL*Net roundtrips to/from client
          2  sorts (memory)
          0  sorts (disk)
         95  rows processed

可以看到，RULE优化器确实根据FROM列表中的位置来选择驱动表，FROM列表中最后一个表被作为驱动表。

但是这只是没有其他优先级更高的规则存在下的情况，只要在表中建立索引，就会导致优先级发生变化，从而导致上面的驱动表选择发生变化：

SQL> CREATE INDEX IND_T1_ID ON T1(ID);

Index created.

SQL> SELECT * 
  2  FROM T2, T1
  3  WHERE T1.ID = T2.ID;

95 rows selected.

Execution Plan
----------------------------------------------------------
Plan hash value: 3280696534

-------------------------------------------------
| Id  | Operation                   | Name      |
-------------------------------------------------
|   0 | SELECT STATEMENT            |           |
|   1 |  TABLE ACCESS BY INDEX ROWID| T1        |
|   2 |   NESTED LOOPS              |           |
|   3 |    TABLE ACCESS FULL        | T2        |
|*  4 |    INDEX RANGE SCAN         | IND_T1_ID |
-------------------------------------------------

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

   4 - access("T1"."ID"="T2"."ID")

Note
-----
   - rule based optimizer used (consider using cbo)

Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
         54  consistent gets
          0  physical reads
          0  redo size
       4545  bytes sent via SQL*Net to client
        558  bytes received via SQL*Net from client
          8  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
         95  rows processed

根据规则T1应该是驱动表，但是建立索引后，索引扫描的优先级远高于全表扫描，因此Oracle选择T2作为驱动表，将T1的执行计划变成了索引扫描。

T2上也存在索引，高优先级的规则对于驱动表的选择上条件等同，这时RULE优化器则又会根据FROM列表来判断驱动表：

SQL> CREATE INDEX IND_T2_ID ON T2(ID);

Index created.

SQL> SELECT *
  2  FROM T2, T1
  3  WHERE T1.ID = T2.ID;

95 rows selected.

Execution Plan
----------------------------------------------------------
Plan hash value: 3621112097

-------------------------------------------------
| Id  | Operation                   | Name      |
-------------------------------------------------
|   0 | SELECT STATEMENT            |           |
|   1 |  TABLE ACCESS BY INDEX ROWID| T2        |
|   2 |   NESTED LOOPS              |           |
|   3 |    TABLE ACCESS FULL        | T1        |
|*  4 |    INDEX RANGE SCAN         | IND_T2_ID |
-------------------------------------------------

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

   4 - access("T1"."ID"="T2"."ID")

Note
-----
   - rule based optimizer used (consider using cbo)

Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
        132  consistent gets
          1  physical reads
          0  redo size
       4545  bytes sent via SQL*Net to client
        558  bytes received via SQL*Net from client
          8  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
         95  rows processed