sql优化_隐式-显示转换

========  测试表1信息   =======
SQL> select count(*) from tb_test;

COUNT(*)
----------
   3000000
  
SQL> select count(*) from tb_test where name='1' or name='0';

COUNT(*)
----------
   3000000

SQL> desc tb_test
 Name                       Null?    Type
 -------------------------- -------- -----------------
 ID                               NUMBER
 INC_DATETIME                       VARCHAR2(19)
 RANDOM_ID                           NUMBER
 RANDOM_STRING                       VARCHAR2(4000)
 NAME                             VARCHAR2(20)

===========  索引信息 ===========

OWNER   TABLE_NAME   INDEX_NAME    INDEX_TYPE
-------  ------------    -------------     ------------
TEST    TB_TEST      IDX_INC              NORMAL
TEST    TB_TEST      NAME_IX              NORMAL
TEST    TB_TEST      ID_IX         　　    NORMAL

====== 测试表2信息 ==========

SQL> desc tb_test1
 Name                                                  Null?    Type
 ----------------------------------------------------- -------- ------------------------------------
 ID                                                           NUMBER
 INC_DATETIME                                                   VARCHAR2(19)
 RANDOM_ID                                                      NUMBER
 RANDOM_STRING                                                   VARCHAR2(4000)

SQL> select count(*) from tb_test1

COUNT(*)
----------
       110

============ 使用隐式转换  ============
使用隐式转换未使用索引，全表扫描，IO开销大，执行速度慢。

explain plan for select   id from tb_test where
id not in ( select id from tb_test1 ) and
RANDOM_STRING=0 and name=1;

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

PLAN_TABLE_OUTPUT
---------------------------------------------------------------------------------
Plan hash value: 2850541703

----------------------------------------------------------------------------------
| Id  | Operation          | Name     | Rows  | Bytes | Cost (%CPU)| Time     |
-----------------------------------------------------------------------------------
|   0 | SELECT STATEMENT   |          |     1 |    26 |  2762   (1)| 00:00:34 |
|*  1 |  HASH JOIN ANTI NA |          |     1 |    26 |  2762   (1)| 00:00:34 |
|*  2 |   TABLE ACCESS FULL| TB_TEST  |     1 |    13 |  2758   (1)| 00:00:34 |
|   3 |   TABLE ACCESS FULL| TB_TEST1 |   110 |  1430 |     4   (0)| 00:00:01 |
-------------------------------------------------------------------------------

===============  使用显示转换 ==================
使用显示转换，索引扫描，IO开销小，执行快。

SQL> explain plan for select   id from tb_test where
  2  id not in ( select id from tb_test1 ) and
  3  RANDOM_STRING='0' and name='1';

Explained.

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

PLAN_TABLE_OUTPUT
-----------------------------------------------------------------------------------------
Plan hash value: 3484870784

-----------------------------------------------------------------------------------------
| Id  | Operation                    | Name     | Rows  | Bytes | Cost (%CPU)| Time     |
---------------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT             |          |     1 |    26 |     8   (0)| 00:00:01 |
|*  1 |  HASH JOIN ANTI NA           |          |     1 |    26 |     8   (0)| 00:00:01 |
|*  2 |   TABLE ACCESS BY INDEX ROWID| TB_TEST  |     1 |    13 |     4   (0)| 00:00:01 |
|*  3 |    INDEX RANGE SCAN          | NAME_IX  |     6 |       |     3   (0)| 00:00:01 |
|   4 |   TABLE ACCESS FULL          | TB_TEST1 |   110 |  1430 |     4   (0)| 00:00:01 |
--------------------------------------------------------------------------------------------------

总结:1、隐式转换发生在字段列上时将使索引失效
     2、RANDOM_STRING和name虽含有大量重复值，使用索引的效率仍然很高。