关于对db_block_gets的理解与实验

实验

一、自己手动创建的小表

创建一个区大小为 40k
SYS@ORCL>show parameter db_block_size

NAME TYPE VALUE
------------------------------------ ----------- ------------------------------
db_block_size integer

SYS@ORCL>create tablespace tyger1 datafile '/u01/app/oracle/oradata/ORCL/tyger1.dbf' size 10m
2 extent management local uniform size 40k;

Tablespace created.

SYS@ORCL>create table test_db1(x int) tablespace tyger1;

Table created.

SYS@ORCL>set autotrace on
SYS@ORCL>insert into test_db1 values(1);

1 row created.

Execution Plan
----------------------------------------------------------

-------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-------------------------------------------------------------------------
| 0 | INSERT STATEMENT | | 1 | 100 | 1 (0)| 00:00:01 |
-------------------------------------------------------------------------

Statistics
----------------------------------------------------------
1 recursive calls
19 db block gets
1 consistent gets
3 physical reads
964 redo size
675 bytes sent via SQL*Net to client
562 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
1 rows processed

SYS@ORCL>insert into test_db1 values(2);

1 row created.

Execution Plan
----------------------------------------------------------

Statistics
----------------------------------------------------------
1 recursive calls
3 db block gets
1 consistent gets
0 physical reads
244 redo size
675 bytes sent via SQL*Net to client
562 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
1 rows processed

2. 创建一个区大小为80k
SYS@ORCL>create tablespace tyger2 datafile '/u01/app/oracle/oradata/ORCL/tyger2.dbf' size 10m
2 extent management local uniform size 80k;

Tablespace created.

SYS@ORCL>create table test_db2(x int) tablespace tyger2;

Table created.

SYS@ORCL>insert into test_db2 values(1);

1 row created.

Execution Plan
----------------------------------------------------------

Statistics
----------------------------------------------------------
1 recursive calls
29 db block gets
1 consistent gets
28 physical reads
1364 redo size
675 bytes sent via SQL*Net to client
562 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
1 rows processed

SYS@ORCL>insert into test_db2 values(2);

1 row created.

Execution Plan
----------------------------------------------------------

Statistics
----------------------------------------------------------
1 recursive calls
3 db block gets
1 consistent gets
0 physical reads
288 redo size
677 bytes sent via SQL*Net to client
562 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
1 rows processed

结论：对于新创建的表来说，因为创建的是空表就没有对表里的空间进行分配，当插入第一
条数据时，就需要对区上的块进行空间分配和对数据字典的一些操作，就会有比较大的db_block_size。如果再次插入数据的话就基本没有对空间的分
配啥的，就会有比较少的db_block_size产生。

所以对于extent指定的区大小来说同样的空表插入同样的数据 db_block_size 可能不同。

对插入更新、删除的实验：
SYS@ORCL>update test_db1 set x=3 where x=1;

1 row updated.

Execution Plan
----------------------------------------------------------
Plan hash value: 2185639234

-------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-------------------------------------------------------------------------------
| 0 | UPDATE STATEMENT | | 1 | 13 | 2 (0)| 00:00:01 |
| 1 | UPDATE | TEST_DB1 | | | | |
|* 2 | TABLE ACCESS FULL| TEST_DB1 | 1 | 13 | 2 (0)| 00:00:01 |
-------------------------------------------------------------------------------

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

2 - filter("X"=1)

Note
-----
- dynamic sampling used for this statement

Statistics
----------------------------------------------------------
28 recursive calls
1 db block gets
11 consistent gets
0 physical reads
388 redo size
678 bytes sent via SQL*Net to client
565 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
1 rows processed

SYS@ORCL>delete test_db1 where x=2;

1 row deleted.

Execution Plan
----------------------------------------------------------
Plan hash value: 3135214910

-------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-------------------------------------------------------------------------------
| 0 | DELETE STATEMENT | | 1 | 13 | 2 (0)| 00:00:01 |
| 1 | DELETE | TEST_DB1 | | | | |
|* 2 | TABLE ACCESS FULL| TEST_DB1 | 1 | 13 | 2 (0)| 00:00:01 |
-------------------------------------------------------------------------------

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

2 - filter("X"=2)

Note
-----
- dynamic sampling used for this statement

Statistics
----------------------------------------------------------
5 recursive calls
1 db block gets
9 consistent gets
0 physical reads
288 redo size
678 bytes sent via SQL*Net to client
557 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
1 rows processed

SYS@ORCL>insert into test_db1 values(&x);
Enter value for x: 1
old 1: insert into test_db1 values(&x)
new 1: insert into test_db1 values(1)

1 row created.

。。。。
SYS@ORCL>commit;

Commit complete.

SYS@ORCL>select * from test_db1;

X
----------
3
1
2
3
4
5
6
7
8
9
19
10
1
11
12
13
14
15
16
17
18

21 rows selected.

SYS@ORCL>alter system flush buffer_cache;

System altered.
SYS@ORCL>update test_db1 set x=21 where x=18;

1 row updated.

Execution Plan
----------------------------------------------------------
Plan hash value: 2185639234

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

2 - filter("X"=18)

Note
-----
- dynamic sampling used for this statement

Statistics
----------------------------------------------------------
5 recursive calls
1 db block gets
9 consistent gets
0 physical reads
412 redo size
678 bytes sent via SQL*Net to client
567 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
1 rows processed

二、对于比较大的表来说

SYS@ORCL>create table test_db1 as select * from dba_objects;

Table created.

SYS@ORCL>insert into test_db1 values('tyger','tyger','tyger',22,23,'tyger','04-SEP-14','04-SEP-14','tyger','t','t','t','t');

1 row created.

Execution Plan
----------------------------------------------------------

Statistics
----------------------------------------------------------
1 recursive calls
15 db block gets
1 consistent gets
5 physical reads
1144 redo size
677 bytes sent via SQL*Net to client
646 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
1 rows processed

SYS@ORCL>alter system flush buffer_cache;

System altered.

SYS@ORCL>update test_db1 set OBJECT_NAME='tom' where owner='tyger';

3 rows updated.

Execution Plan
----------------------------------------------------------
Plan hash value: 2185639234

-------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-------------------------------------------------------------------------------
| 0 | UPDATE STATEMENT | | 8 | 664 | 154 (2)| 00:00:02 |
| 1 | UPDATE | TEST_DB1 | | | | |
|* 2 | TABLE ACCESS FULL| TEST_DB1 | 8 | 664 | 154 (2)| 00:00:02 |
-------------------------------------------------------------------------------

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

2 - filter("OWNER"='tyger')

Note
-----
- dynamic sampling used for this statement

Statistics
----------------------------------------------------------
5 recursive calls
3 db block gets
769 consistent gets
687 physical reads
824 redo size
679 bytes sent via SQL*Net to client
589 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
3 rows processed
SYS@ORCL>delete test_db1 where owner='tyger';

3 rows deleted.

Execution Plan
----------------------------------------------------------
Plan hash value: 3135214910

-------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
-------------------------------------------------------------------------------
| 0 | DELETE STATEMENT | | 8 | 136 | 154 (2)| 00:00:02 |
| 1 | DELETE | TEST_DB1 | | | | |
|* 2 | TABLE ACCESS FULL| TEST_DB1 | 8 | 136 | 154 (2)| 00:00:02 |
-------------------------------------------------------------------------------

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

2 - filter("OWNER"='tyger')

Note
-----
- dynamic sampling used for this statement

Statistics
----------------------------------------------------------
4 recursive calls
3 db block gets
769 consistent gets
0 physical reads
1064 redo size
679 bytes sent via SQL*Net to client
567 bytes received via SQL*Net from client
4 SQL*Net roundtrips to/from client
1 sorts (memory)
0 sorts (disk)
3 rows processed

结论：对于占用多个段的大表来说，可能对数据修改时对数据字典或者对于区、块的分配都包含在 physical reads中。

感想：

对于生产库来说，这个值一般不会太考虑到底数字是怎么来的，因为数字都比较大，一般只在乎它的大小数量级。

关于对db_block_gets的理解与实验的更多相关文章

关于Jmeter线程数Ramp-Up.循环次数的理解和实验数据
1. 关于线程组参数线程组:即一个线程组实例里面包括多个串行的请求或动作.一个线程组的从启动到结束的时间取决于你线程中的步骤数量. 线程数:即用户数,在Ramp-up时间内(包括循环),简单把线程数 ...
setInterval的简单理解和实验
setInterval的用法 setInterval(fn_name,time_num); setInterval(fn_name,time_num,这里是函数参数); 意思是,现在不执行fn_nam ...
2分钟理解文件IO -我对文件IO的理解与实验对比
本文介绍了不同的IO方式以及他们之间的效率比较 1.一次读取写入单个字节(读取400M的文件浪费了很久,等了很久没读取完成,证明其效率很差) public class CopyFileDemo { p ...
《深入理解计算机系统》实验三 —— Buf Lab
这是CSAPP的第三个实验,主要让我们熟悉GDB的使用,理解程序栈帧的结构和缓冲区溢出的原理. 实验目的本实验的目的在于加深对IA-32函数调用规则和栈结构的具体理解.实验的主要内容是对一个可执 ...
北京电子科技学院（BESTI）实验报告5
北京电子科技学院(BESTI)实验报告5 课程: 信息安全系统设计基础班级:1452.1453 姓名:(按贡献大小排名) 郑凯杰.周恩德学号:(按贡献大小排名) 20145314.20145217 ...
linux内核分析作业8：理解进程调度时机跟踪分析进程调度与进程切换的过程
1. 实验目的选择一个系统调用(13号系统调用time除外),系统调用列表,使用库函数API和C代码中嵌入汇编代码两种方式使用同一个系统调用分析汇编代码调用系统调用的工作过程,特别是参数的传递的方 ...
20145209&20145309信息安全系统设计基础实验报告（3）
实验内容.步骤与体会: 实验过程的理解,实验指导书中知识点的理解. (1)为什么在双击了GIVEIO和JTAG驱动安装文件后还需要手动进行配置? 因为安装文件只是将驱动文件释放了出来,并没有在系统中将 ...
Linux第三次实验报告
北京电子科技学院(BESTI) 实验报告课程:信息安全系统设计基础班级:201352 姓名:池彬宁贺邦学号:20135212 2013520 ...
信息安全系统设计基础实验一 20135210&20135218
北京电子科技学院(BESTI) 实验报告课程: 密码系统设计基础 ...

随机推荐

linux-阿里云ECS部署PPTP（centos)
请参考以下步骤:(centos6.5中测试通过) 1.服务器端安装软件 1.1 首先安装ppp,命令: [root@test ~]#yum install -y ppp 提示Complete! ,表示 ...
python django第一天
django第一天,上一次是使用了virtualenv(沙盒),这一次就直接用virtualenv直接建一个django的环境,用来自己练着试试django: source django/Script ...
Java二叉搜索树实现
树集合了数组(查找速度快)和链表(插入.删除速度快)的优点二叉树是一种特殊的树,即:树中的每个节点最多只能有两个子节点二叉搜索树是一种特殊的二叉树,即:节点的左子节点的值都小于这个节点的值,节点的 ...
如何定位到append的当前位置，不用拉滚动条scrollIntoView方法
var bb_mes_con = $('bb_mes_con'); var mes_html = document.createElement('div'); mes_html.setAttribut ...
POJ 1159 - Palindrome (LCS，滚动数组)
Palindrome Time Limit: 3000MS Memory Limit: 65536K Total Submissions: 55018 Accepted: 19024 Desc ...
poj 3544 Journey with Pigs
Journey with Pigs Time Limit: 1000MS Memory Limit: 65536K Total Submissions: 3004 Accepted: 922 ...
【转载】LinkedIn是如何优化Kafka的
http://www.wtoutiao.com/p/18d5RY0.html 在LinkedIn的数据基础设施中,Kafka是核心支柱之一.来自LinkedIn的工程师曾经就Kafka写过一系列的专题 ...
Android手机播放电脑视频文件-屌丝必备
今天早上一到办公室,照常打开博客园看文章,看到有一片文章是用 http://www.cnblogs.com/wdfrog/p/3738180.html 看到这哥们实现的方法好复杂,又是配置电脑端,又 ...
两种查看eclipse或MyEclipse是64bit还是32bit的方法
方法一: 对于:eclipse 打开eclipse后:Help-->About Eclipse-->Installation Details-->Configuration 如果看到 ...
ORA-00001: unique constraint (...) violated解决方案
ORA-00001: unique constraint (...) violated 的解决方案今天往Oracle数据库里插入数据一条记录的时候,报错了, 控制台抛出异常:违反唯一性约定, 我以为 ...

关于对db_block_gets的理解与实验

关于对db_block_gets的理解与实验的更多相关文章

随机推荐

热门专题