daf0 2014-10-08
一、实验说明:
操作系统:rhel 5.4 x86
数据库:Oracle 11g R2
实验说明:该实验是为了说明B-Tree索引性能优于BitMap索引的情况。
二、实验操作:
首先创建一张t_btree表,并建立B-Tree索引,索引键是object_id:
SQL> create table t_btree as select * from dba_objects;
Table created.
SQL> create index ind_tree on t_btree(object_id);
Index created.
执行两次下面的查询语句,并显示执行计划:
SQL> set autotrace traceonly;
SQL> select * from t_btree where object_id=9899;
Execution Plan
----------------------------------------------------------
Plan hash value: 447474086
----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 207 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T_BTREE | 1 | 207 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IND_TREE | 1 | | 1 (0)| 00:00:01 |
----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("OBJECT_ID"=9899)
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
312 recursive calls
0 db block gets
108 consistent gets
289 physical reads
0 redo size
1404 bytes sent via SQL*Net to client
419 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 * from t_btree where object_id=9899;
Execution Plan
----------------------------------------------------------
Plan hash value: 447474086
----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 207 | 2 (0)| 00:00:01 |
| 1 | TABLE ACCESS BY INDEX ROWID| T_BTREE | 1 | 207 | 2 (0)| 00:00:01 |
|* 2 | INDEX RANGE SCAN | IND_TREE | 1 | | 1 (0)| 00:00:01 |
----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
2 - access("OBJECT_ID"=9899)
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
0 recursive calls
0 db block gets
4 consistent gets
0 physical reads
0 redo size
1404 bytes sent via SQL*Net to client
419 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
同样执行之前的语句两次:
SQL> select * from t_bmap where object_id=9899;
Execution Plan
----------------------------------------------------------
Plan hash value: 3763176822
----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 207 | 110 (0)| 00:00:02 |
| 1 | TABLE ACCESS BY INDEX ROWID | T_BMAP | 1 | 207 | 110 (0)| 00:00:02 |
| 2 | BITMAP CONVERSION TO ROWIDS| | | | | |
|* 3 | BITMAP INDEX SINGLE VALUE | IND_MAP | | | | |
----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
3 - access("OBJECT_ID"=9899)
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
312 recursive calls
0 db block gets
98 consistent gets
266 physical reads
0 redo size
1404 bytes sent via SQL*Net to client
419 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 * from t_bmap where object_id=9899;
Execution Plan
----------------------------------------------------------
Plan hash value: 3763176822
----------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 207 | 110 (0)| 00:00:02 |
| 1 | TABLE ACCESS BY INDEX ROWID | T_BMAP | 1 | 207 | 110 (0)| 00:00:02 |
| 2 | BITMAP CONVERSION TO ROWIDS| | | | | |
|* 3 | BITMAP INDEX SINGLE VALUE | IND_MAP | | | | |
----------------------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
3 - access("OBJECT_ID"=9899)
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
7 recursive calls
0 db block gets
68 consistent gets
0 physical reads
0 redo size
1404 bytes sent via SQL*Net to client
419 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
从上面的查询中,我们可以得到,分别给两张内容一样的表做查询的时候,在执行第二次的时候是属于软解析:
从一致性读上比较,B-Tree索引的consistent gets是4,BitMap的是68;
从Cost的消耗上看,B-Tree索引的COST是2,而BitMap的是110。
在索引键是主键或者唯一性约束的情况下B-Tree索引的效率要优于BitMap索引。