内容简介:
1.普通B-TREE 索引;
2.唯一B-TREE 索引;
3.复合索引;
ORACLE 默认的索引类型为B-TREE 索引,表中的行标识符(ROWID)和行相关的列值被存储在一个平衡树的树状结构的索引块中;使用B-TREE索引有以下几个原因:
▢ 提高SQL语句的性能;
▢ 强制执行主键和唯一键约束的唯一性;
▢ 减少通过主键和外键约束关联的父表和子表间潜在的锁定问题 ;
1.普通B-TREE 索引
在一张未建立任何索引的500万行人员信息表中根据人员ID查询人员信息
select id,name,gender,homeaddr from th01 where id=998698;
--------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 38 | 16715 (1)| 00:03:21 |
|* 1 | TABLE ACCESS FULL| TH01 | 1 | 38 | 16715 (1)| 00:03:21 |
--------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter("ID"=998698)
Statistics
----------------------------------------------------------
61222 consistent gets
61208 physical reads
通过观察执行计划,CBO优化器执行了全表扫描,一致读取61222个块,61208个物理读,基于性能的考虑和表结构的分析,为其B-TREE索引:
SQL> CREATE INDEX IND_TH01_ID ON TH01(ID) TABLESPACE TBS02;
Index created.
Elapsed: 00:00:33.03
SQL> execute dbms_stats.gather_table_stats('sywu','th01',cascade=>true);
PL/SQL procedure successfully completed.
Elapsed: 00:00:04.17
SQL> @/oracle/getind
TABLE_NAME INDEX_NAME COLUMN_NAME SIZE_GB INDEX_TY STATUS LOGGING DEGREE NUM_ROWS DISTINCT_KEYS
------------------------------ ------------------------------ ------------------------------ ---------- -------- -------- --------- ---------- -------- ---------- ------------- ---
TH01 IND_TH01_ID ID .091796875 NORMAL VALID YES 1 DISABLED 5000000 5000000
显然对于高基数的列创建B-TREE索引是明智之选,对表进行分析后再次查询:
SQL>select id,name,gender,homeaddr from th01 where id=998698;
Elapsed: 00:00:00.00
-------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 38 | 4 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| TH01 | 1 | 38 | 4 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IND_TH01_ID | 1 | | 3 (0)| 00:00:01 |
------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("ID"=998698)
Statistics
----------------------------------------------------------
5 consistent gets
0 physical reads
基于成本的考虑,CBO优化器选择了通过索引的方式读取数据,一致读取5个块,有效减少了额外的物理读;做个基于索引列的统计查询:
SQL> select count(id) from th01;
Elapsed: 00:00:00.15
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 6 | 3170 (1)| 00:00:39 |
| 1 | SORT AGGREGATE | | 1 | 6 | | |
| 2 | INDEX FAST FULL SCAN| IND_TH01_ID | 5000K| 28M| 3170 (1)| 00:00:39 |
-------------------------------------------------------------------------------------
Statistics
----------------------------------------------------------
11810 consistent gets
11794 physical reads
CBO 优化器选择了全索引扫描,依旧消耗额外的资源;但当统计列发生改变时:
SQL> select count(*) from th01;
Elapsed: 00:00:00.14
-------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 16707 (1)| 00:03:21 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | TABLE ACCESS FULL| TH01 | 5000K| 16707 (1)| 00:03:21 |
Statistics
----------------------------------------------------------
61221 consistent gets
61208 physical reads
此时CBO 优化器选择了全表扫描,并消耗更多的资源;
2.唯一B-TREE 索引:
在为表创建(主键、唯一约束)时,ORACLE 会默认创建一个B-TREE索引,这样既保证了数据的唯一性也提高了数据的检索效率:
SQL> alter table th01 add constraints cs_th01_uq unique(idcard);
Table altered.
Elapsed: 00:00:56.11
TABLE_NAME INDEX_NAME COLUMN_NAME SIZE_GB INDEX_TY STATUS LOGGING DEGREE COMPRESS NUM_ROWS DISTINCT_KEYS
------------------------------ ------------------------------ ------------------------------ ---------- -------- -------- --------- ---------- -------- ---------- ------------- ---
TH01 CS_TH01_UQ IDCARD .15625 NORMAL VALID YES 1 DISABLED 4969898 4969898
以IDCARD查询人员信息:
------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 35 | 3 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| TH01 | 1 | 35 | 3 (0)| 00:00:01 |
|* 2 | INDEX UNIQUE SCAN | CS_TH01_UQ | 1 | | 2 (0)| 00:00:01 |
------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("IDCARD"='562456864646565545')
Statistics
----------------------------------------------------------
3 consistent gets
0 physical reads
做为默认创建的索引,它依旧能高效的工作,但ORACLE是不允许将其作为独立的索引删除的,只能通过删除约束的方式删除;对于主键,它的情况要复杂些,因为还要考虑外键的约束;基于这种方式创建的索引,当约束被删除时还要重新创建索引,显然在一张大表上花费的代价和时间是昂贵的; so,采用如下的方式合理的建立约束和索引:
SQL> alter table th01 add constraints CS_TH01_UQ unique(idcard)
2* using index tablespace tbs03 ;
Table altered.
Elapsed: 00:00:59.27
倘若有一天业务发生了改变,唯一约束已经不是必须的,但索引是必须的,那只需要删除约束保留索引:
SQL> alter table th01 drop constraints CS_TH01_UQ keep index;
Table altered.
Elapsed: 00:00:00.01
再次通过IDCARD 查询人员信息:
SQL> select id,name,idcard from th01 where idcard='56234256878945';
------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 35 | 3 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| TH01 | 1 | 35 | 3 (0)| 00:00:01 |
|* 2 | INDEX UNIQUE SCAN | CS_TH01_UQ | 1 | | 2 (0)| 00:00:01 |
------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("IDCARD"='56234256878945')
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
3 consistent gets
只创建唯一索引:
SQL> create unique index ind_th02 on th02(idcard) tablespace tbs03;
唯一索引与唯一约束相比,唯一索引只创建索引而不添加约束,它保证索引列数值唯一性,允许有空值;
3.复合索引:
可以在多个列上创建索引,其结果称为复合索引或组合索引:
SQL> create index ind_th01_union on th01(id,name,idcard) tablespace tbs03;
当查询的WHERE子句引用了索引的所有列或者只是前导列,CBO会使用复合索引
SQL>select id,name,idcard from th01 where idcard='9876534655635666' and id=68956254 and name='张三';
-----------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-----------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 35 | 3 (0)| 00:00:01 |
|* 1 | INDEX RANGE SCAN| IND_TH01_UNION | 1 | 35 | 3 (0)| 00:00:01 |
-----------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - access("ID"=68956254 AND "NAME"='张三' AND
"IDCARD"='9876534655635666')
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
3 consistent gets
2 physical reads
通过查询结果,对于之前创建的单列索引(IND_TH01_ID、CS_TH01_UQ),优化器已不再使用;对于复合索引( IND_TH01_UNION) 来说, id | id, name| id,name,idcard 三个组合都被认为是前导列,假如我只是在WHERE 子句中引用了第一个主导列ID,那么优化器依旧会选择复合索引(IND_TH01_UNION )忽略单列索引( IND_TH01_ID)
SQL> select * from th01 where id=698698;
----------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 77 | 4 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| TH01 | 1 | 77 | 4 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IND_TH01_UNION | 1 | | 3 (0)| 00:00:01 |
----------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("ID"=698698)
Statistics
----------------------------------------------------------
0 db block gets
5 consistent gets
假如where 子句中的条件不符合复合索引前导列的要求,那么优化器会忽略复合索引( IND_TH01_UNION)选择单列索引(CS_TH01_UQ):
SQL> select * from th01 where idcard='5623546566564665';
------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 77 | 3 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| TH01 | 1 | 77 | 3 (0)| 00:00:01 |
|* 2 | INDEX UNIQUE SCAN | CS_TH01_UQ | 1 | | 2 (0)| 00:00:01 |
------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("IDCARD"='5623546566564665')
Statistics
----------------------------------------------------------
1 recursive calls
0 db block gets
3 consistent gets
1 physical reads
对于此时的查询条件 where idcard='5623546566564665' 已经不再符合( id | id, name| id,name,idcard) 复合索引前导列的条件,优化器选择单列索引( CS_TH01_UQ);又假如,我的WHERE 子句条件符合复合索引前导列要求但不是全部满足:
SQL> select * from th01 where id=698698 and name='张三';
----------------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 77 | 4 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| TH01 | 1 | 77 | 4 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IND_TH01_UNION | 1 | | 3 (0)| 00:00:01 |
----------------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("ID"=698698 AND "NAME"='张三')
Statistics
----------------------------------------------------------
3 consistent gets
0 physical reads
通过分析(where id=698698 and name='张三' )符合复合索引前导列要求,优化器选择复合索引( IND_TH01_UNION) 忽略单列索引( IND_TH01_ID);
创建复合索引时,排序是个很大的问题,ORACLE 建议将最频繁访问的列放在索引中最靠前的位置,应避免使用低基数的列作为复合索引的前导列.