ITPub博客

首页 > Linux操作系统 > Linux操作系统 > ORACLE的表分析策略[转]

ORACLE的表分析策略[转]

原创 Linux操作系统 作者:sunyudong 时间:2012-06-08 12:37:01 0 删除 编辑

ORACLE的表分析策略[转]

    对表进行分析,通常情况下可以对表,索引,列进行单独分析,或者进行组合分析,但这三者哪些是相对重要的,哪些分析显得不那么重要?通过本篇文章的实验相信大家也会对直方图有更一步的了解.
1.首先创建测试表,并插入100000条数据
SQL> create table test(id number,nick varchar2(30));

Table created.

SQL> begin
  2      for i in 1..100000 loop
  3            insert into test(id) values(i);
  4      end loop;
  5      commit;
  6  end;
  7  /

PL/SQL procedure successfully completed.

更新nick字段,使数据发生严重倾斜
SQL> update test set nick='abc' where rownum<99999;

99998 rows updated.

SQL> commit;

Commit complete.

SQL> create index idx_test_nick on test(nick);

Index created.


SQL> update test set nick='def' where nick is null;

2 rows updated.

SQL> commit;

Commit complete.

--只对索引进行分析
SQL> analyze index idx_test_nick compute statistics;

Index analyzed.

SQL> select index_name,LEAF_BLOCKS,DISTINCT_KEYS,NUM_ROWS from user_indexes where index_name='IDX_TEST_NICK';

INDEX_NAME                     LEAF_BLOCKS DISTINCT_KEYS   NUM_ROWS
------------------------------ ----------- ------------- ----------
IDX_TEST_NICK                          210             2     100000


SQL> select COLUMN_NAME,NUM_BUCKETS,num_distinct from USER_tab_columns where table_name='TEST';

COLUMN_NAME                    NUM_BUCKETS NUM_DISTINCT
------------------------------ ----------- ------------
ID
NICK

查看只分析索引的情况下的执行计划,从执行计划可以看出,优化器选择RBO,都走索引
SQL> set autotrace trace exp
SQL> select * from test where nick ='abc';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE
   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST'
   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_NICK' (NON-UNIQUE)

 

SQL> select * from test where nick ='def';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE
   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST'
   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_NICK' (NON-UNIQUE)


分析有数据严重倾斜的nick列后,也没有产生正确的执行计划,此时使用的优化器仍然是RBO
通过此实验,说明只分析索引和列,ORACLE不会使用CBO的优化器
SQL> analyze table test compute statistics for columns size 2 nick;

Table analyzed.

SQL> select * from test where nick ='abc';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE
   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST'
   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_NICK' (NON-UNIQUE)

 

SQL> select * from test where nick ='def';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE
   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST'
   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_NICK' (NON-UNIQUE)


现在分析表后,产生了正确的执行计划
SQL> analyze table test compute statistics for table;

Table analyzed.

SQL> select * from test where nick ='abc';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE (Cost=50 Card=99998 Bytes=
          1499970)

   1    0   TABLE ACCESS (FULL) OF 'TEST' (Cost=50 Card=99998 Bytes=14
          99970)


SQL> select * from test where nick ='def';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE (Cost=1 Card=2 Bytes=30)
   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST' (Cost=1 Card=2 Byt
          es=30)

   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_NICK' (NON-UNIQUE) (Cost
          =1 Card=2)
        
删除所有的统计数据,并只对表与列进行分析,不分析索引,ORACLE使用CBO的优化器,并产生了正确的执行计划
SQL> analyze table test delete statistics;

Table analyzed.

SQL> analyze table test compute statistics for table for columns size 2 nick;

Table analyzed.

SQL> select * from test where nick ='abc';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE (Cost=50 Card=99998 Bytes=
          1499970)

   1    0   TABLE ACCESS (FULL) OF 'TEST' (Cost=50 Card=99998 Bytes=14
          99970)

SQL> select * from test where nick ='def';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE (Cost=1 Card=2 Bytes=30)
   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST' (Cost=1 Card=2 Byt
          es=30)

   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_NICK' (NON-UNIQUE) (Cost
          =1 Card=2)


创建TEST表ID列上的索引,但不对索引进行分析
SQL> create index idx_test_id on test(id);

Index created.

SQL> select * from test where id=1;

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE (Cost=1 Card=1000 Bytes=15
          000)

   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST' (Cost=1 Card=1000
          Bytes=15000)

   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_ID' (NON-UNIQUE) (Cost=1
           Card=400)

当条件中即有id,又有nick时,因为nick上有直方图,ORACLE知道nick='abc'的值特别的多,所以不走IDX_TEST_NICK索引,走IDX_TEST_ID上的索引
SQL> select * from test where id=5 and nick='abc';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE (Cost=1 Card=1000 Bytes=15
          000)

   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST' (Cost=1 Card=1000
          Bytes=15000)

   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_ID' (NON-UNIQUE) (Cost=1
           Card=400)

当条件中即有id,又有nick时,因为nick上有直方图,ORACLE知道nick='def'的值特别的少,所以走IDX_TEST_NICK上的索引,不走IDX_TEST_ID索引


SQL> select * from test where id=5 and nick='def';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE (Cost=1 Card=1 Bytes=15)
   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST' (Cost=1 Card=1 Byt
          es=15)

   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_NICK' (NON-UNIQUE) (Cost
          =1 Card=2)

在分析ID列后,ORACLE发现ID列的选择度更高,所以不再选择IDX_TEST_NICK索引,而是选择IDX_TEST_ID
SQL> analyze table test compute statistics for columns size 1 id;

Table analyzed.

SQL> select * from test where id=5 and nick='def';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE (Cost=1 Card=1 Bytes=7)
   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST' (Cost=1 Card=1 Byt
          es=7)

   2    1     INDEX (RANGE SCAN) OF 'IDX_TEST_ID' (NON-UNIQUE) (Cost=1
           Card=1)

下面来看另外一种情况,我们删除所有的统计数据,然后在ID列上创建唯一索引,在此条件下,只分析表与分析列nick,我们看到ORACLE走了正 确的执行计划,走了UK_TEST_ID,其实从这里也给我们带来很多的启示:在主键与唯一键约束的列上是否需要直方图的问题?如果在这些列上有像这样的 查询where id > 100 and id < 1000,我们还是需要有直方图的,但除此之外,好像真的没有直方图的必要了!
SQL> analyze table test delete statistics;  

Table analyzed.

SQL> drop index idx_test_id;

Index dropped.

SQL> create unique index uk_test_id on test(id);

Index created.

SQL> analyze table test compute statistics for table for columns size 2 nick;

Table analyzed.

SQL> select * from test where id=5 and nick='def';

Execution Plan
----------------------------------------------------------
   0      SELECT STATEMENT ptimizer=CHOOSE (Cost=1 Card=1 Bytes=15)
   1    0   TABLE ACCESS (BY INDEX ROWID) OF 'TEST' (Cost=1 Card=1 Byt
          es=15)

   2    1     INDEX (UNIQUE SCAN) OF 'UK_TEST_ID' (UNIQUE) (Cost=1 Car
          d=1)

从以上一系列的实验可以看出,对ORACLE的优化器CBO来说,表的分析与列的分析才是最重要的,索引的分析次之。还有我们可以考虑我们的哪些列 上需要直方图,对于bucket的个数问题,oracle的默认值是75个,所以根据你的应用规则,选择合适的桶数对性能也是有帮助的。因为不必要的桶的 个数的大量增加,必然会带来SQL语句硬解析时产生执行计划的复杂度问题。

 

来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/312949/viewspace-732269/,如需转载,请注明出处,否则将追究法律责任。

请登录后发表评论 登录
全部评论

注册时间:2011-04-27

  • 博文量
    23
  • 访问量
    44582