oracle索引失效总结

--oracle索引失效
--创建测试表
create table t(
empno number,
ename varchar2(20),
deptno number
);
--创建索引
create index idx_deptno on t(deptno);
--创建复合索引
create index idx_empno_deptno on t(empno,deptno);
drop index idx_empno_deptno;
create index idx_empno_deptno on t(deptno,empno);
--插入测试数据
insert into t values(1001,'A',10);

1    <> 
SQL> select * from t where deptno=10;

Execution Plan
----------------------------------------------------------
Plan hash value: 3776569808

------------------------------------------------------------------------------------------
| Id  | Operation            | Name     | Rows  | Bytes | Cost (%CPU)| Time     |
------------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT        |         |     2 |    76 |     1   (0)| 00:00:01 |
|   1 |  TABLE ACCESS BY INDEX ROWID| T      |     2 |    76 |     1   (0)| 00:00:01 |
|*  2 |   INDEX RANGE SCAN        | IDX_DEPTNO |     2 |     |     1   (0)| 00:00:01 |
------------------------------------------------------------------------------------------


SQL> select * from t where deptno<>10;

Execution Plan
----------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation      | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |     |     7 |   266 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T     |     7 |   266 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

   
2 模糊查询 like '%_' 
SQL> select * from t where deptno like '%_1';     

Execution Plan
----------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation      | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |     |     1 |    38 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T     |     1 |    38 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

3 单独引用复合索引里非第一位置的索引列.
SQL> select * from t where empno=1005;

Execution Plan
----------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation      | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |     |     1 |    38 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T     |     1 |    38 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

4 字符型字段为数字时在where条件里不添加引号
alter table t add id varchar2(20);
create index idx_id on t(id);
SQL> select * from t where id=33;

Execution Plan
----------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation      | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |     |     1 |    50 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T     |     1 |    50 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

SQL> select * from t where id='33';

Execution Plan
----------------------------------------------------------
Plan hash value: 827754323

--------------------------------------------------------------------------------------
| Id  | Operation            | Name   | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT        |         |       1 |      50 |       2   (0)| 00:00:01 |
|   1 |  TABLE ACCESS BY INDEX ROWID| T      |       1 |      50 |       2   (0)| 00:00:01 |
|*  2 |   INDEX RANGE SCAN        | IDX_ID |       1 |         |       1   (0)| 00:00:01 |
--------------------------------------------------------------------------------------

5 对索引列进行运算.需要建立函数索引
SQL> select * from t where deptno*2=20;

Execution Plan
----------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation      | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |     |     2 |   100 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T     |     2 |   100 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

SQL> select * from t where substr(deptno,1,1)=3;

Execution Plan
----------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation      | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |     |     1 |    50 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T     |     1 |    50 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

6 not in ,not exist.
SQL> select * from t where deptno not in ('10','20','30','40','50','60');

Execution Plan
----------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation      | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |     |     3 |   150 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T     |     3 |   150 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

7 基于cost成本分析(oracle因为走全表成本会更小):查询小表,或者返回值大概在10%以上
8 B-tree索引 is null不会走,is not null会走
SQL> select * from t where id is null;

Execution Plan
----------------------------------------------------------
Plan hash value: 1601196873

--------------------------------------------------------------------------
| Id  | Operation      | Name | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------
|   0 | SELECT STATEMENT  |     |     7 |   350 |     3   (0)| 00:00:01 |
|*  1 |  TABLE ACCESS FULL| T     |     7 |   350 |     3   (0)| 00:00:01 |
--------------------------------------------------------------------------

SQL> select * from t where id is not null;

Execution Plan
----------------------------------------------------------
Plan hash value: 875909553

--------------------------------------------------------------------------------------
| Id  | Operation            | Name   | Rows  | Bytes | Cost (%CPU)| Time     |
--------------------------------------------------------------------------------------
|   0 | SELECT STATEMENT        |         |       3 |     150 |       2   (0)| 00:00:01 |
|   1 |  TABLE ACCESS BY INDEX ROWID| T      |       3 |     150 |       2   (0)| 00:00:01 |
|*  2 |   INDEX FULL SCAN        | IDX_ID |      10 |         |       1   (0)| 00:00:01 |
--------------------------------------------------------------------------------------

索引失效总结:
1.没有查询条件,查询条件没有建立索引
2.查询条件没有使用索引引导列
3.基于cost优化器,查询结果集30%以上
4.索引本身失效
5.查询条件使用函数在索引列上
6.隐式转换
7.模糊查询
8.not in , null ,<> 

 

posted @ 2018-10-10 21:16  wangzihong  阅读(1176)  评论(0编辑  收藏  举报