您可以在针对视图的查询中使用提示来强制 Oracle 使用基表上的索引。但是您需要知道基础视图中基表的别名(如果有)。一般语法是/*+ index(<<alias of view from query>> <<alias of table from view>> <<index name>>) */
一个例子
1) 创建一个包含 10,000 个相同行的表,并在该表上创建索引。索引没有选择性,所以 Oracle 不想使用它
SQL> ed
Wrote file afiedt.buf
1 create table foo
2 as
3 select 1 col1
4 from dual
5* connect by level <= 10000
SQL> /
Table created.
SQL> create index idx_foo on foo(col1);
Index created.
2) 验证索引没有正常使用但是Oracle会提示使用
SQL> set autotrace traceonly;
SQL> select * from foo where col1 = 1;
10000 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 1245013993
--------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 10000 | 126K| 7 (0)| 00:00:01 |
|* 1 | TABLE ACCESS FULL| FOO | 10000 | 126K| 7 (0)| 00:00:01 |
--------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter("COL1"=1)
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
9 recursive calls
0 db block gets
713 consistent gets
5 physical reads
0 redo size
172444 bytes sent via SQL*Net to client
7849 bytes received via SQL*Net from client
668 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
10000 rows processed
SQL> select /*+ index(foo idx_foo) */ *
2 from foo
3 where col1 = 1;
10000 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 15880034
----------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 10000 | 126K| 25 (0)| 00:00:01 |
|* 1 | INDEX RANGE SCAN| IDX_FOO | 10000 | 126K| 25 (0)| 00:00:01 |
----------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - access("COL1"=1)
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
7 recursive calls
0 db block gets
715 consistent gets
15 physical reads
0 redo size
172444 bytes sent via SQL*Net to client
7849 bytes received via SQL*Net from client
668 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
10000 rows processed
3) 现在创建视图。通过指定查询中的视图别名和视图定义中的表别名来验证针对视图的正常查询不使用索引,而是强制使用索引
SQL> create view vw_foo
2 as
3 select col1
4 from foo f;
View created.
SQL> select col1
2 from vw_foo
3 where col1 = 1;
10000 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 1245013993
--------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
--------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 10000 | 126K| 7 (0)| 00:00:01 |
|* 1 | TABLE ACCESS FULL| FOO | 10000 | 126K| 7 (0)| 00:00:01 |
--------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - filter("COL1"=1)
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
16 recursive calls
0 db block gets
715 consistent gets
0 physical reads
0 redo size
172444 bytes sent via SQL*Net to client
7849 bytes received via SQL*Net from client
668 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
10000 rows processed
SQL> select /*+ index(vf f idx_foo) */ col1
2 from vw_foo vf
3 where col1 = 1;
10000 rows selected.
Execution Plan
----------------------------------------------------------
Plan hash value: 15880034
----------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
----------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 10000 | 126K| 25 (0)| 00:00:01 |
|* 1 | INDEX RANGE SCAN| IDX_FOO | 10000 | 126K| 25 (0)| 00:00:01 |
----------------------------------------------------------------------------
Predicate Information (identified by operation id):
---------------------------------------------------
1 - access("COL1"=1)
Note
-----
- dynamic sampling used for this statement (level=2)
Statistics
----------------------------------------------------------
14 recursive calls
0 db block gets
717 consistent gets
0 physical reads
0 redo size
172444 bytes sent via SQL*Net to client
7849 bytes received via SQL*Net from client
668 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
10000 rows processed
SQL>
尽管如此,在尝试调整查询时,提示通常是最后的手段——通常最好找出优化器缺少哪些信息并提供适当的统计信息,以便它可以自己做出正确的选择。这是一个更稳定的解决方案。更重要的是,当您减少到指定涉及多层别名的提示时 - 例如,触摸视图定义的人很容易通过更改表名的别名来破坏您的查询。