ITPub博客

首页 > Linux操作系统 > Linux操作系统 > 系统优化之---分区表加速查询速度和删除速度

系统优化之---分区表加速查询速度和删除速度

原创 Linux操作系统 作者:Aminiy 时间:2013-06-05 21:37:12 0 删除 编辑
1、删除表里的数据最用下面语句
alter table t_name drop partition p_name;
alter table t_name truncate partition p_name;
delete 方式弊端:消耗大量的系统资源和无法释放空间

SQL> create user test identified by test account unlock; 
User created.
SQL> grant dba to test;
Grant succeeded.
SQL> grant resource,connect to test;
Grant succeeded.
SQL> conn test/test as sysdba
Connected.
SQL>  create table t as select object_id,object_name from dba_objects;
create table t as select object_id,object_name from dba_objects
Table created.
SQL> exec dbms_stats.gather_table_stats(user,'t');

PL/SQL procedure successfully completed.

SQL> set autotrace trace exp stat;
SQL> delete from t where object_id <10000;

9708 rows deleted.


Execution Plan
----------------------------------------------------------
Plan hash value: 3335594643

---------------------------------------------------------------------------
| Id  | Operation          | Name | Rows  | Bytes | Cost (%CPU)| Time     |
---------------------------------------------------------------------------
|   0 | DELETE STATEMENT   |      |  9715 | 48575 |    96   (2)| 00:00:02 |
|   1 |  DELETE            | T    |       |       |            |          |
|*  2 |   TABLE ACCESS FULL| T    |  9715 | 48575 |    96   (2)| 00:00:02 |
---------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   2 - filter("OBJECT_ID"<10000)


Statistics
----------------------------------------------------------
        200  recursive calls
      10106  db block gets
        373  consistent gets
          0  physical reads
    2609544  redo size
        678  bytes sent via SQL*Net to client
        608  bytes received via SQL*Net from client
          3  SQL*Net roundtrips to/from client
          5  sorts (memory)
          0  sorts (disk)
       9708  rows processed
这种情况下,delete操作产生了10106+373条数据块(db block gets+consistent gets)同是产生了2.6M的redo日志。
即使用创建索引的方法也无法避免资源消耗
SQL> rollback;

Rollback complete.

SQL> create index ind_t on t(object_id);

Index created.
SQL> exec dbms_stats.gather_index_stats(user,'ind_t')

PL/SQL procedure successfully completed.

SQL>  delete from t where object_id <10000;

9708 rows deleted.


Execution Plan
----------------------------------------------------------
Plan hash value: 3974964266

---------------------------------------------------------------------------
| Id  | Operation         | Name  | Rows  | Bytes | Cost (%CPU)| Time     |
---------------------------------------------------------------------------
|   0 | DELETE STATEMENT  |       |  9715 | 48575 |    23   (0)| 00:00:01 |
|   1 |  DELETE           | T     |       |       |            |          |
|*  2 |   INDEX RANGE SCAN| IND_T |  9715 | 48575 |    23   (0)| 00:00:01 |
---------------------------------------------------------------------------

Predicate Information (identified by operation id):
---------------------------------------------------

   2 - access("OBJECT_ID"<10000)


Statistics
----------------------------------------------------------
        282  recursive calls
      10370  db block gets
         90  consistent gets
          0  physical reads
    2777076  redo size
        680  bytes sent via SQL*Net to client
        609  bytes received via SQL*Net from client
          3  SQL*Net roundtrips to/from client
          8  sorts (memory)
          0  sorts (disk)
       9708  rows processed
上面两种情况消耗的资源差不多都是1000多个数据块,相反使用truncate和drop 消耗的资源小的多。

  create table t1(object_id int,object_name varchar2(1000)) partition by range(object_id)
 (partition p1 values less than(10000),
partition p2 values less than(20000),
partition p3 values less than(30000),
partition p4 values less than(40000)
,partition pm values less than(maxvalue));

 create table t1(object_id int,object_name varchar2(1000)) partition by range(object_id) (partition p1 values less than(10000),partition p2 values less than(20000),partition p3 values less than(30000),partition p4 values less than(40000),partition pm values less than(maxvalue));
SQL> insert into t1 select * from t;

62751 rows created.


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

--------------------------------------------------------------------------------
-

| Id  | Operation                | Name | Rows  | Bytes | Cost (%CPU)| Time
|

--------------------------------------------------------------------------------
-

|   0 | INSERT STATEMENT         |      | 72459 |  2052K|    95   (0)| 00:00:02
|

|   1 |  LOAD TABLE CONVENTIONAL | T1   |       |       |            |
|

|   2 |   TABLE ACCESS FULL      | T    | 72459 |  2052K|    95   (0)| 00:00:02
|

--------------------------------------------------------------------------------
-



Statistics
----------------------------------------------------------
       1170  recursive calls
      39860  db block gets
      38041  consistent gets
          0  physical reads
   10866864  redo size
        680  bytes sent via SQL*Net to client
        602  bytes received via SQL*Net from client
          3  SQL*Net roundtrips to/from client
          1  sorts (memory)
          0  sorts (disk)
      62751  rows processed


SQL> commit;
启动SQL TRACE实用工具
对会话启动SQL TRACE 
SQL> alter session set sql_trace=true;

Session altered.
SQL> alter table t1 truncate partition p1;

SQL> truncate table t1;

Table truncated.

SQL> insert into t1 select * from t;

62751 rows created.

SQL> commit;

Commit complete.

SQL> alter session set sql_trace=true;

Session altered.

SQL> alter table t1 drop partition p1;

Table altered.
关闭SQL TRACE实用工具
对会话关闭SQL TRACE 
SQL> alter session set sql_trace=false;

Session altered.
获得进程信息,选择需要跟踪的进程:
select s.USERNAME,s.SID,s.SERIAL#,s.COMMAND from v$session s where s.USERNAME='TEST';

USERNAME                              SID    SERIAL#    COMMAND
------------------------------ ---------- ---------- ----------
TEST                                   33         52          2

SQL> show user
USER is "SYS"
设置跟踪:
SQL>  exec sys.dbms_system.set_sql_trace_in_session(33,52,true)

PL/SQL procedure successfully completed.

可以等候片刻,跟踪session执行任务,捕获sql操作…
停止跟踪:
SQL> exec sys.dbms_system.set_sql_trace_in_session(33,52,false);

PL/SQL procedure successfully completed.

 tkorof工具 分析ORACLE跟踪文件并且产生一个更加人性化清晰的输出结果的可执行工具
[oracle@haoxy trace]$ pwd
/u01/app/diag/rdbms/hxy/hxy/trace
[oracle@haoxy trace]$ tkprof hxy_ora_10283.trc tra.txt print=600 record=sql.txt sys=no
vi tra.txt

alter table t1 truncate partition p1


call     count       cpu    elapsed       disk      query    current        rows
------- ------  -------- ---------- ---------- ---------- ----------  ----------
Parse        1      0.00       0.06          0          4          1           0
Execute      1      0.03       0.11          5          1         82           0
Fetch        0      0.00       0.00          0          0          0           0
------- ------  -------- ---------- ---------- ---------- ----------  ----------
total        2      0.04       0.18          5          5         83           0

Misses in library cache during parse: 1
Optimizer mode: ALL_ROWS
Parsing user id: 91

alter table t1 drop partition p1


call     count       cpu    elapsed       disk      query    current        rows
------- ------  -------- ---------- ---------- ---------- ----------  ----------
Parse        1      0.02       0.12          0          0          1           0
Execute      1      0.04       0.11          1          1         35           0
Fetch        0      0.00       0.00          0          0          0           0
------- ------  -------- ---------- ---------- ---------- ----------  ----------
total        2      0.06       0.23          1          1         36           0

Misses in library cache during parse: 1
Optimizer mode: ALL_ROWS
Parsing user id: 91
由此实验可以看出DDL消耗资源远小于DML资源,DDl消耗的资源是对数据字典的修改,这个值基本是不变的,DML是随着数据量的增大消耗的资源也随之增大。

来自 “ ITPUB博客 ” ,链接:http://blog.itpub.net/24492954/viewspace-765656/,如需转载,请注明出处,否则将追究法律责任。

上一篇: rman备份(转载)
下一篇: 10046事件(转)
请登录后发表评论 登录
全部评论

注册时间:2011-04-09

  • 博文量
    41
  • 访问量
    239032