使用索引的误区之二：使用了 和 != 操作符，导致查询不使用索引

使用索引的误区之二：使用了 <> 和 != 操作符，导致查询不使用索引

http://blog.csdn.net/lunar2000/archive/2004/11/24/193475.aspx

首先，请记住这个结论：[/b]

使用了<> 和!=后，就不会使用索引[/b]
[/b]
例如，下面的例子使用了<>，所以查询没有用到索引[/b]
select empno from emp where empno <>10;

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
--------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost |
--------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | |
|* 1 | TABLE ACCESS FULL | EMP | | | |
--------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter("EMP"."EMPNO"<>10)
Note: rule based optimization

14 rows selected

[/b]
将上面的查条件“empno <>10”转换成“empno <10 and empno>10”后，就可以使用索引了[/b]
select empno from emp where empno <10 and empno>10;

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
--------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost |
--------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | |
|* 1 | INDEX RANGE SCAN | EMP_ID1 | | | |
--------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - access("EMP"."EMPNO">10 AND "EMP"."EMPNO"<10)
Note: rule based optimization

14 rows selected

SQL>

再看下面的例子：[/b]
由于使用了前导列，所以使用了索引，后面的"!="是从索引范围扫描的结果中筛选合适的记录的[/b]
select empno from emp where empno <=10 and ename != 'RICH';

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
--------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost |
--------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | |
|* 1 | INDEX RANGE SCAN | EMP_ID1 | | | |
--------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - access("EMP"."EMPNO"<=10)
filter("EMP"."EMPNO"<=10 AND "EMP"."ENAME"<>'RICH')
Note: rule based optimization

15 rows selected


再做一个试验：[/b]
SQL> desc dept
Name Type Nullable Default Comments
------ ------------ -------- ------- --------
DEPTNO NUMBER(2) Y
DNAME VARCHAR2(14) Y
LOC VARCHAR2(13) Y

创建一个单键索引：
SQL> create index dept_id1 on dept(dname);

Index created

如果使用"<>"，则查询不使用索引：[/b]
select depTno from dept where dname <> 'DEVELOPER';

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
--------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost |
--------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | |
|* 1 | TABLE ACCESS FULL | DEPT | | | |
--------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter("DEPT"."DNAME"<>'DEVELOPER')
Note: rule based optimization

14 rows selected

将条件修改为“dname <'DEVELOPER' and dname>'DEVELOPER'”，则可以使用索引[/b]
select deptno from dept where dname <'DEVELOPER' and dname>'DEVELOPER';

PLAN_TABLE_OUTPUT
--------------------------------------------------------------------------------
---------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost |
---------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | | | |
| 1 | TABLE ACCESS BY INDEX ROWID| DEPT | | | |
|* 2 | INDEX RANGE SCAN | DEPT_ID1 | | | |
---------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("DEPT"."DNAME">'DEVELOPER' AND "DEPT"."DNAME"<'DEVELOPER')
Note: rule based optimization

15 rows selected

SQL>

评论 #piner 发表于2005-01-20 17:15:00 IP: 218.108.88.* 首先，请记住这个结论： 使用了<> 和!=后，就不会使用索引 ????? 这个结论不正确，有误导倾向 #lunar 发表于2005-03-01 16:38:00 IP: 218.249.251.* 是有些问题，或许应该改成，如果使用了<> 和 !=操作，如果不用提示的话，系统就不用索引，这样应该可以了吧，呵呵。 #ccq 发表于2006-03-23 21:17:00 IP: 172.17.35.* select empno from emp where empno <>10; 等价于: select empno from emp where empno <10 and empno>10; 不会吧? 是否可以调整为: select empno from emp where empno <10 union all select empno from emp where empno >10; #lunar 发表于2006-03-24 08:21:00 IP: 218.206.69.* 对，粗心了，不好意思 ：）