在9i中因为某个执行计划因为Oracle Peeking绑定变量的控制导致现有的执行计划不正确,需要清除掉这条sql语句的执行计划.在10g中提供了dbms_shared_pool.purge(见:清除掉shared pool中某条sql语句方法),但是在9i中未提供好的方法,一般来说可以通过对相关表的DDL操作,收集统计信息,授权操作可以实现清除对于表执行计划.注:这些操作不会只清空特定SQL执行计划,而是会清除该表相关的所有执行计划,所以操作需要慎重(影响肯定比flush shared_pool小)
模拟测试数据

SQL> create table t_xifenfei (id number,name varchar2(100));
Table created.
SQL> insert into t_xifenfei values(1,'www.xifenfei.com');
1 row created.
SQL> commit;

清除执行计划1:修改表结构

SQL>  alter system flush shared_pool;
System altered.
SQL> select * from t_xifenfei;
        ID NAME
---------- -------------------
         1 www.xifenfei.com
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL> select OPERATION from v$sql_plan where hash_value=1067507827;
OPERATION
------------------------------------------------------------
SELECT STATEMENT
TABLE ACCESS
SQL> alter table t_xifenfei  add fei varchar2(10);
Table altered.
SQL> alter table t_xifenfei drop COLUMN fei;
Table altered.
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL>  select count(*) from v$sql_plan where hash_value=1067507827;
  COUNT(*)
----------
         0

清除执行计划2:重新收集统计信息

--DBMS_STATS收集统计信息
SQL> alter system flush shared_pool;
System altered.
SQL> select * from t_xifenfei;
        ID NAME
---------- -------------------
         1 www.xifenfei.com
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL> select OPERATION from v$sql_plan where hash_value=1067507827;
OPERATION
------------------------------------------------------------
SELECT STATEMENT
TABLE ACCESS
SQL> EXEC DBMS_STATS.gather_table_stats(user,'T_XIFENFEI');
PL/SQL procedure successfully completed.
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL> select OPERATION from v$sql_plan where hash_value=1067507827;
no rows selected
--analyze收集统计信息(不推荐)
SQL> alter system flush shared_pool;
System altered.
SQL> select * from t_xifenfei;
        ID NAME
---------- -------------------
         1 www.xifenfei.com
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL> select OPERATION from v$sql_plan where hash_value=1067507827;
OPERATION
------------------------------------------------------------
SELECT STATEMENT
TABLE ACCESS
SQL> analyze table  t_xifenfei compute statistics;
Table analyzed.
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL> select OPERATION from v$sql_plan where hash_value=1067507827;
no rows selected

清除执行计划3:创建INDEX

SQL> alter system flush shared_pool;
System altered.
SQL> select * from t_xifenfei;
        ID NAME
---------- -------------------
         1 www.xifenfei.com
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL>  select OPERATION from v$sql_plan where hash_value=1067507827;
OPERATION
------------------------------------------------------------
SELECT STATEMENT
TABLE ACCESS
SQL> create index i_txifenfei on t_xifenfei(id) online;
Index created.
SQL> drop index i_txifenfei ;
Index dropped.
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL> select OPERATION from v$sql_plan where hash_value=1067507827;
no rows selected

清除执行计划3:GRANT/REVOKE操作

SQL> alter system flush shared_pool;
System altered.
SQL> select * from t_xifenfei;
        ID NAME
---------- -------------------
         1 www.xifenfei.com
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL> select OPERATION from v$sql_plan where hash_value=1067507827;
OPERATION
------------------------------------------------------------
SELECT STATEMENT
TABLE ACCESS
SQL> GRANT SELECT ON T_XIFENFEI TO SYSTEM;
Grant succeeded.
SQL> select sql_text,HASH_VALUE from v$sqlAREA where sql_text like '%xifenfei';
SQL_TEXT                                           HASH_VALUE
-------------------------------------------------- ----------
select * from t_xifenfei                           1067507827
SQL> select OPERATION from v$sql_plan where hash_value=1067507827;
no rows selected

近期有朋友对于单个表上的index各种情况比较模糊,这里对于单个表上,单个index出现的大多数情况进行了总结性测试,给出了测试结果,至于为什么出现这样的试验结果未做过多解释,给读者留下思考的空间.本篇文章仅仅是为了测试hint对index的影响,而不是说明走各种index方式的好坏.参考: INDEX FULL SCAN vs INDEX FAST FULL SCAN
创建表模拟测试

SQL> create table t_xifenfei as select object_id,object_name from dba_objects;
Table created.
SQL>  create index i_t_object_id on t_xifenfei(object_id);
Index created.
SQL> exec dbms_stats.gather_table_stats(USER,'T_XIFENFEI',cascade=>true);
PL/SQL procedure successfully completed.
SQL> desc t_xifenfei
 Name                                      Null?    Type
 ----------------------------------------- -------- ----------------------------
 OBJECT_ID                                          NUMBER
 OBJECT_NAME                                        VARCHAR2(128)

TABLE ACCESS FULL

SQL> SET AUTOT TRACE EXP STAT
SQL> SELECT OBJECT_ID FROM T_XIFENFEI;
49838 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 548923532
--------------------------------------------------------------------------------
| Id  | Operation         | Name       | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |            | 49838 |   243K|    57   (2)| 00:00:01 |
|   1 |  TABLE ACCESS FULL| T_XIFENFEI | 49838 |   243K|    57   (2)| 00:00:01 |
--------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
       3544  consistent gets
          0  physical reads
          0  redo size
     721203  bytes sent via SQL*Net to client
      36927  bytes received via SQL*Net from client
       3324  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
      49838  rows processed
SQL> SELECT /*+ INDEX(T i_t_object_id) */ OBJECT_ID FROM T_XIFENFEI;
49838 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 548923532
--------------------------------------------------------------------------------
| Id  | Operation         | Name       | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |            | 49838 |   243K|    57   (2)| 00:00:01 |
|   1 |  TABLE ACCESS FULL| T_XIFENFEI | 49838 |   243K|    57   (2)| 00:00:01 |
--------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
       3544  consistent gets
          0  physical reads
          0  redo size
     721203  bytes sent via SQL*Net to client
      36927  bytes received via SQL*Net from client
       3324  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
      49838  rows processed

从上面的执行计划中可知,此时走了全表扫描. 由于我们需要查询的列为object_id,因此理论上只需要读取索引就应该可以返回所有数据,而此时为什么是全表扫描呢？ 这是因为NULL值与索引的特性所决定的.即null值不会被存储到B树索引.因此应该为表 t_xifenfei 的列 object_id 添加 not null 约束.

INDEX FAST FULL SCAN

SQL> alter table t_xifenfei modify(object_id not null);
Table altered.
SQL> SELECT  object_id from t_xifenfei;
49838 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 2036340805
--------------------------------------------------------------------------------------
| Id  | Operation            | Name          | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT     |               | 49838 |   243K|    27   (4)| 00:00:01 |
|   1 |  INDEX FAST FULL SCAN| I_T_OBJECT_ID | 49838 |   243K|    27   (4)| 00:00:01 |
--------------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
       3432  consistent gets
          0  physical reads
          0  redo size
     721203  bytes sent via SQL*Net to client
      36927  bytes received via SQL*Net from client
       3324  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
      49838  rows processed

INDEX FAST FULL SCAN:当在高速缓存中没有找到所需的索引块时,则根据db_file_multiblock_read_count的值进行多块读操作.对于索引的分支结构只是简单的获取,然后扫描所有的叶结点.其结果是导致索引结构没有访问,获取的数据没有根据索引键的顺序排序.INDEX FAST FULL SCAN使用multiblock_read,故产生db file scattered reads 事件.

INDEX RANGE SCAN

SQL> select object_id from t_xifenfei where object_id<10;
8 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 2197008162
----------------------------------------------------------------------------------
| Id  | Operation        | Name          | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------------
|   0 | SELECT STATEMENT |               |     2 |    10 |     2   (0)| 00:00:01 |
|*  1 |  INDEX RANGE SCAN| I_T_OBJECT_ID |     2 |    10 |     2   (0)| 00:00:01 |
----------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
   1 - access("OBJECT_ID"<10)
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
          3  consistent gets
          0  physical reads
          0  redo size
        499  bytes sent via SQL*Net to client
        385  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
          8  rows processed
SQL> select /*+ index_ffs(t i_t_object_id) */ object_id from t_xifenfei t where object_id<10;
8 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 2036340805
--------------------------------------------------------------------------------------
| Id  | Operation            | Name          | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT     |               |     2 |    10 |    27   (4)| 00:00:01 |
|*  1 |  INDEX FAST FULL SCAN| I_T_OBJECT_ID |     2 |    10 |    27   (4)| 00:00:01 |
--------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
   1 - filter("OBJECT_ID"<10)
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
        118  consistent gets
          0  physical reads
          0  redo size
        499  bytes sent via SQL*Net to client
        385  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
          8  rows processed

这里可以看出index_ffs已经生效,但是对于这样的情况hint index_ffs效率一般来说不会太高.
<br>
<strong>INDEX FULL SCAN</strong>

SQL> SELECT /*+ INDEX(T i_t_object_id) */ object_id  from t_xifenfei t;
49838 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 431110666
----------------------------------------------------------------------------------
| Id  | Operation        | Name          | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------------
|   0 | SELECT STATEMENT |               | 49838 |   243K|   113   (2)| 00:00:02 |
|   1 |  INDEX FULL SCAN | I_T_OBJECT_ID | 49838 |   243K|   113   (2)| 00:00:02 |
----------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
       3426  consistent gets
          0  physical reads
          0  redo size
     721203  bytes sent via SQL*Net to client
      36927  bytes received via SQL*Net from client
       3324  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
      49838  rows processed

INDEX FULL SCAN:完全按照索引存储的顺序依次访问整个索引树.当访问到叶结点之后,按照双向链表方式读取相连节点的值.换言之,对于索引上所有的数据是按照有序的方式来读取的.如果索引块没有在高速缓存中被找到时,则需要从数据文件中单块进行读取.对于需要读取大量数据的全索引扫描而言,这将使其变得低效.INDEX FULL SCAN使用single read,故产生db file sequential reads事件.新版的Oracle支持db file parallel reads方式.
HINT INDEX不会使用INDEX FAST FULL SCAN功能.

INDEX列ORDER BY

SQL> SELECT OBJECT_ID FROM T_XIFENFEI order by object_id ;
49838 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 431110666
----------------------------------------------------------------------------------
| Id  | Operation        | Name          | Rows  | Bytes | Cost (%CPU)| Time     |
----------------------------------------------------------------------------------
|   0 | SELECT STATEMENT |               | 49838 |   243K|   113   (2)| 00:00:02 |
|   1 |  INDEX FULL SCAN | I_T_OBJECT_ID | 49838 |   243K|   113   (2)| 00:00:02 |
----------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
       3426  consistent gets
          0  physical reads
          0  redo size
     721203  bytes sent via SQL*Net to client
      36927  bytes received via SQL*Net from client
       3324  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
      49838  rows processed
SQL> SELECT OBJECT_ID FROM T_XIFENFEI order by object_id  desc;
49838 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 2808014233
--------------------------------------------------------------------------------------------
| Id  | Operation                  | Name          | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT           |               | 49838 |   243K|   113   (2)| 00:00:02 |
|   1 |  INDEX FULL SCAN DESCENDING| I_T_OBJECT_ID | 49838 |   243K|   113   (2)| 00:00:02 |
--------------------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
       3427  consistent gets
          0  physical reads
          0  redo size
     721203  bytes sent via SQL*Net to client
      36927  bytes received via SQL*Net from client
       3324  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
      49838  rows processed
SQL> SELECT  /*+ index_ffs(t i_t_object_id) */ object_id from t_xifenfei t order by object_id;
49838 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 2527678987
-----------------------------------------------------------------------------------------------
| Id  | Operation             | Name          | Rows  | Bytes |TempSpc| Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT      |               | 49838 |   243K|       |   185   (4)| 00:00:03 |
|   1 |  SORT ORDER BY        |               | 49838 |   243K|  1192K|   185   (4)| 00:00:03 |
|   2 |   INDEX FAST FULL SCAN| I_T_OBJECT_ID | 49838 |   243K|       |    27   (4)| 00:00:01 |
-----------------------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
        117  consistent gets
          0  physical reads
          0  redo size
     721203  bytes sent via SQL*Net to client
      36927  bytes received via SQL*Net from client
       3324  SQL*Net roundtrips to/from client
          1  sorts (memory)
          0  sorts (disk)
      49838  rows processed

对于index 列排序,默认情况下会使用INDEX FULL SCAN/INDEX FULL SCAN DESCENDING而不选择使用INDEX FAST FULL SCAN,因为INDEX FAST FULL SCAN获得数据后,还需要做一次SORT ORDER BY操作

INDEX FAST FULL SCAN+SORT AGGREGATE

SQL> SELECT  count(object_id) FROM T_XIFENFEI;
Execution Plan
----------------------------------------------------------
Plan hash value: 3095383276
-------------------------------------------------------------------------------
| Id  | Operation             | Name          | Rows  | Cost (%CPU)| Time     |
-------------------------------------------------------------------------------
|   0 | SELECT STATEMENT      |               |     1 |    27   (4)| 00:00:01 |
|   1 |  SORT AGGREGATE       |               |     1 |            |          |
|   2 |   INDEX FAST FULL SCAN| I_T_OBJECT_ID | 49838 |    27   (4)| 00:00:01 |
-------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
        117  consistent gets
          0  physical reads
          0  redo size
        421  bytes sent via SQL*Net to client
        385  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
          1  rows processed
SQL> SELECT  /*+ INDEX(T i_t_object_id) */ count(object_id) FROM T_XIFENFEI t;
Execution Plan
----------------------------------------------------------
Plan hash value: 3079973526
--------------------------------------------------------------------------
| Id  | Operation        | Name          | Rows  | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT |               |     1 |   113   (2)| 00:00:02 |
|   1 |  SORT AGGREGATE  |               |     1 |            |          |
|   2 |   INDEX FULL SCAN| I_T_OBJECT_ID | 49838 |   113   (2)| 00:00:02 |
--------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
        111  consistent gets
          0  physical reads
          0  redo size
        421  bytes sent via SQL*Net to client
        385  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
          1  rows processed

sort aggregate通常发生在使用一些聚合函数的时候,sum(),avg(),min(),max(),count()等等,实际上sort aggregate不做真正的sort,并不会用到排序空间,而是通过一个全局变量＋全表或全索引扫描来实现.这样的操作在默认情况下使用INDEX FAST FULL SCAN

INDEX FULL SCAN (MIN/MAX)

SQL> SELECT  max(object_id) FROM T_XIFENFEI;
Execution Plan
----------------------------------------------------------
Plan hash value: 2939893782
--------------------------------------------------------------------------------------------
| Id  | Operation                  | Name          | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT           |               |     1 |     5 |     2   (0)| 00:00:01 |
|   1 |  SORT AGGREGATE            |               |     1 |     5 |            |          |
|   2 |   INDEX FULL SCAN (MIN/MAX)| I_T_OBJECT_ID | 49838 |   243K|     2   (0)| 00:00:01 |
--------------------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
          2  consistent gets
          0  physical reads
          0  redo size
        419  bytes sent via SQL*Net to client
        385  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
          1  rows processed
SQL>  SELECT /*+ index_ffs(t i_t_object_id) */ max(object_id) FROM T_XIFENFEI t;
Execution Plan
----------------------------------------------------------
Plan hash value: 2939893782
--------------------------------------------------------------------------------------------
| Id  | Operation                  | Name          | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT           |               |     1 |     5 |    27   (4)| 00:00:01 |
|   1 |  SORT AGGREGATE            |               |     1 |     5 |            |          |
|   2 |   INDEX FULL SCAN (MIN/MAX)| I_T_OBJECT_ID | 49838 |   243K|    27   (4)| 00:00:01 |
--------------------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
          1  recursive calls
          0  db block gets
          2  consistent gets
          0  physical reads
          0  redo size
        419  bytes sent via SQL*Net to client
        385  bytes received via SQL*Net from client
          2  SQL*Net roundtrips to/from client
          0  sorts (memory)
          0  sorts (disk)
          1  rows processed

对于这样的查询INDEX FULL SCAN (MIN/MAX)明显是最优,但是此处奇怪的是使用了index_ffs提示无效,如果有知道的朋友,麻烦告知原因.