count(*)、count(val)和count(1)的解释
一、关于count的一些谣言：
1、count(*)比count(val)更慢！项目组必须用count(val)，不准用count(*)，谁用扣谁钱！
2、count(*)用不到索引，count(val)才能用到。
3、count(*)是统计出全表的记录，是吞吐量的操作，肯定用不到索引。
4、count(1)比count(*)的速度快。
二、验证count(*)和count(val)
1、首先创建一个表，使用count(*)和count(val)查询比较：
----删除echo表----
SQL> drop table echo purge;
drop table echo purge
*
第 1 行出现错误:
ORA-00942: 表或视图不存在 ----创建一张echo的测试表----
SQL> create table echo as select * from dba_objects; 表已创建。 SQL> update echo set object_id = rownum; 已更新72509行。 SQL> commit; 提交完成。 SQL> set timing on
SQL> set linesize 100
SQL> set autotrace on
SQL> select count(*) from echo; COUNT(*)
----------
72509 已用时间: 00: 00: 00.01 执行计划
----------------------------------------------------------
Plan hash value: 99109176 -------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 290 (1)| 00:00:04 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | TABLE ACCESS FULL| ECHO | 80064 | 290 (1)| 00:00:04 |
------------------------------------------------------------------- Note
-----
- dynamic sampling used for this statement (level=2) 统计信息
----------------------------------------------------------
4 recursive calls
0 db block gets
1265 consistent gets
0 physical reads
11060 redo size
425 bytes sent via SQL*Net to client
415 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 count(*) from echo; COUNT(*)
----------
72509 已用时间: 00: 00: 00.01 执行计划
----------------------------------------------------------
Plan hash value: 99109176 -------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 290 (1)| 00:00:04 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | TABLE ACCESS FULL| ECHO | 80064 | 290 (1)| 00:00:04 |
------------------------------------------------------------------- Note
-----
- dynamic sampling used for this statement (level=2) 统计信息
----------------------------------------------------------
0 recursive calls
0 db block gets
1038 consistent gets
0 physical reads
0 redo size
425 bytes sent via SQL*Net to client
415 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 count(object_id) from echo; COUNT(OBJECT_ID)
----------------
72509 已用时间: 00: 00: 00.01 执行计划
----------------------------------------------------------
Plan hash value: 99109176 ---------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 13 | 290 (1)| 00:00:04 |
| 1 | SORT AGGREGATE | | 1 | 13 | | |
| 2 | TABLE ACCESS FULL| ECHO | 80064 | 1016K| 290 (1)| 00:00:04 |
--------------------------------------------------------------------------- Note
-----
- dynamic sampling used for this statement (level=2) 统计信息
----------------------------------------------------------
4 recursive calls
0 db block gets
1112 consistent gets
0 physical reads
0 redo size
433 bytes sent via SQL*Net to client
415 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 count(object_id) from echo; COUNT(OBJECT_ID)
----------------
72509 已用时间: 00: 00: 00.01 执行计划
----------------------------------------------------------
Plan hash value: 99109176 ---------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 13 | 290 (1)| 00:00:04 |
| 1 | SORT AGGREGATE | | 1 | 13 | | |
| 2 | TABLE ACCESS FULL| ECHO | 80064 | 1016K| 290 (1)| 00:00:04 |
--------------------------------------------------------------------------- Note
-----
- dynamic sampling used for this statement (level=2) 统计信息
----------------------------------------------------------
0 recursive calls
0 db block gets
1038 consistent gets
0 physical reads
0 redo size
433 bytes sent via SQL*Net to client
415 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
从上面的执行计划可以看出count(*)和count(val)是一样快的。
2、建立索引做比较
SQL> create index idx_object_id on echo(object_id); 索引已创建。 已用时间: 00: 00: 05.69
SQL> select count(*) from echo; COUNT(*)
----------
72509 已用时间: 00: 00: 00.05 执行计划
----------------------------------------------------------
Plan hash value: 99109176 -------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 290 (1)| 00:00:04 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | TABLE ACCESS FULL| ECHO | 80064 | 290 (1)| 00:00:04 |
------------------------------------------------------------------- Note
-----
- dynamic sampling used for this statement (level=2) 统计信息
----------------------------------------------------------
5 recursive calls
0 db block gets
1113 consistent gets
0 physical reads
0 redo size
425 bytes sent via SQL*Net to client
415 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 count(object_id) from echo; COUNT(OBJECT_ID)
----------------
72509 已用时间: 00: 00: 00.08 执行计划
----------------------------------------------------------
Plan hash value: 1131838604 ---------------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time |
---------------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 13 | 49 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | 13 | | |
| 2 | INDEX FAST FULL SCAN| IDX_OBJECT_ID | 80064 | 1016K| 49 (0)| 00:00:01 |
--------------------------------------------------------------------------------------- Note
-----
- dynamic sampling used for this statement (level=2) 统计信息
----------------------------------------------------------
4 recursive calls
0 db block gets
244 consistent gets
161 physical reads
0 redo size
433 bytes sent via SQL*Net to client
415 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
哇，原来真的是用count(val)比count(*)要快啊，因为count(*)不能用到索引，而count(val)可以，真相真是如此吗？
3、将object_id设置为非空
SQL> alter table echo modify object_id not null; 表已更改。 已用时间: 00: 00: 01.41 SQL> select count(*) from echo; COUNT(*)
----------
72509 已用时间: 00: 00: 00.00 执行计划
----------------------------------------------------------
Plan hash value: 1131838604 -------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 49 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | INDEX FAST FULL SCAN| IDX_OBJECT_ID | 80064 | 49 (0)| 00:00:01 |
------------------------------------------------------------------------------- Note
-----
- dynamic sampling used for this statement (level=2) 统计信息
----------------------------------------------------------
0 recursive calls
0 db block gets
169 consistent gets
0 physical reads
0 redo size
425 bytes sent via SQL*Net to client
415 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 count(object_id) from echo; COUNT(OBJECT_ID)
----------------
72509 已用时间: 00: 00: 00.01 执行计划
----------------------------------------------------------
Plan hash value: 1131838604 -------------------------------------------------------------------------------
| Id | Operation | Name | Rows | Cost (%CPU)| Time |
-------------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 49 (0)| 00:00:01 |
| 1 | SORT AGGREGATE | | 1 | | |
| 2 | INDEX FAST FULL SCAN| IDX_OBJECT_ID | 80064 | 49 (0)| 00:00:01 |
------------------------------------------------------------------------------- Note
-----
- dynamic sampling used for this statement (level=2) 统计信息
----------------------------------------------------------
0 recursive calls
0 db block gets
169 consistent gets
0 physical reads
0 redo size
433 bytes sent via SQL*Net to client
415 bytes received via SQL*Net from client
2 SQL*Net roundtrips to/from client
0 sorts (memory)
0 sorts (disk)
1 rows processed
--看来count(val)和count(*)其实一样快，如果索引列是非空的，count(*)可用到索引，此时一样快！真相真是如此吗？
其实两者根本没有可比性，性能比较首先考虑写法等价，这两个语句根本就不等价。
结论：
其实优化器里的算法是这么玩的，列的偏移量决定性能，列越靠后，访问的开销越大。由于count(*)的算法与列偏移量无关，所以count(*)最快，count(最后列val)最慢。
3、用实验验证上面的结论：
SQL> set serveroutput on
SQL> set echo on
SQL> drop table t purge; 表已删除。
----构造出有25个字段的表T----
DECLARE
l_sql VARCHAR2(32767);
BEGIN
l_sql := 'CREATE TABLE t (';
FOR i IN 1..25
LOOP
l_sql := l_sql || 'n' || i || ' NUMBER,';
END LOOP;
l_sql := l_sql || 'pad VARCHAR2(1000)) PCTFREE 10';
EXECUTE IMMEDIATE l_sql;
END;
12 / PL/SQL 过程已成功完成。
----将记录还有这个表T中填充----
DECLARE
l_sql VARCHAR2(32767);
BEGIN
l_sql := 'INSERT INTO t SELECT ';
FOR i IN 1..25
LOOP
l_sql := l_sql || '0,';
END LOOP;
l_sql := l_sql || 'NULL FROM dual CONNECT BY level <= 10000';
EXECUTE IMMEDIATE l_sql;
COMMIT;
END;
13 / PL/SQL 过程已成功完成。 execute dbms_stats.gather_table_stats(ownname=>user, tabname=>'t')
SELECT num_rows, blocks FROM user_tables WHERE table_name = 'T'; PL/SQL 过程已成功完成。 SQL>
NUM_ROWS BLOCKS
---------- ----------
10000 80
----以下动作观察执行速度，比较发现count(*)最快，count(最大列)最慢----
SQL> DECLARE
l_dummy PLS_INTEGER;
l_start PLS_INTEGER;
l_stop PLS_INTEGER;
l_sql VARCHAR2(100);
BEGIN
l_start := dbms_utility.get_time;
FOR j IN 1..1000
LOOP
EXECUTE IMMEDIATE 'SELECT count(*) FROM t' INTO l_dummy;
END LOOP;
l_stop := dbms_utility.get_time;
dbms_output.put_line((l_stop-l_start)/100);
14
FOR i IN 1..25
LOOP
l_sql := 'SELECT count(n' || i || ') FROM t';
l_start := dbms_utility.get_time;
FOR j IN 1..1000
LOOP
EXECUTE IMMEDIATE l_sql INTO l_dummy;
END LOOP;
l_stop := dbms_utility.get_time;
dbms_output.put_line((l_stop-l_start)/100);
END LOOP;
END;
27 /
.18
.33
.39
.38
.42
.4
.45
.49
.48
.46
.48
.48
.55
.51
.56
.57
.61
.62
.75
.67
.68
.7
.73
.78
.77
.81
PL/SQL 过程已成功完成。
三、验证count(*)和count(1)
沿用TOM大师的解释：
count(*)、count(val)和count(1)的解释的更多相关文章
- SQLSERVER 里SELECT COUNT(1) 和SELECT COUNT(*)哪个性能好？
SQLSERVER 里SELECT COUNT(1) 和SELECT COUNT(*)哪个性能好? 今天遇到某人在我以前写的一篇文章里问到 如果统计信息没来得及更新的话,那岂不是统计出来的数据时错误的 ...
- COUNT(*),count(1),COUNT(ALL expression)，COUNT(DISTINCT expression)
创建一个测试表 IF OBJECT_ID( 'dbo.T1' , 'U' )IS NOT NULL BEGIN DROP TABLE dbo.T1; END; GO )); GO INSERT INT ...
- hql中不能写count(1)能够写count(a.id)
hql中不能写count(1)能够写count(a.id)里面写详细的属性 String hql="select new com.haiyisoft.vo.entity.cc.repo.Bu ...
- count（*）、count（1）和count（列名）的区别
count(*).count(1)和count(列名)的区别 1.执行效果上: l count(*)包括了所有的列,相当于行数,在统计结果的时候,不会忽略列值为NULL l count(1)包 ...
- 今天犯了个小错误：_dataArray.count>1 和_dataArray.count>0搞混淆了
_dataArray.count>1 和_dataArray.count>0搞混淆了:当数据为一条时,条件不成立.应该_dataArray.count>＝1 或者>0 ( ...
- COUNT(*)、COUNT(主键)、COUNT(1)
MyISAM引擎,记录数是结构的一部分,已存cache在内存中; InnoDB引擎,需要重新计算,id是主键的话,会加快扫描速度: 所以select count(*) MyISAM完胜! MyISA ...
- 【优化】COUNT(1)、COUNT(*)、COUNT(常量)、COUNT(主键)、COUNT(ROWID)、COUNT(非空列)、COUNT(允许为空列)、COUNT(DISTINCT 列名)
[优化]COUNT(1).COUNT(*).COUNT(常量).COUNT(主键).COUNT(ROWID).COUNT(非空列).COUNT(允许为空列).COUNT(DISTINCT 列名) 1. ...
- oracle count（*） 和count（列）性能
一直以为oracle中count(列)比count(*) 快,这篇文件解释了一下: http://blog.csdn.net/szstephenzhou/article/details/8446481
- COUNT(*),count(1),COUNT(ALL expression)，COUNT(DISTINCT expression) BY Group by
select column_2,count(column_2) as 'count(column_2)' ,count(column_1) as 'count(column_1)' ,count(*) ...
随机推荐
- RabbitMQ学习总结 第六篇：Topic类型的exchange
目录 RabbitMQ学习总结 第一篇:理论篇 RabbitMQ学习总结 第二篇:快速入门HelloWorld RabbitMQ学习总结 第三篇:工作队列Work Queue RabbitMQ学习总结 ...
- MEF 松耦合
MEF天生就有解耦合的特性,虽然它不是为解耦而生,而主要是为插件类应用的开发而设计.如果主要是为了解除耦合的话可以使用IoC,Unity等. Unity 微软的IOC 代码: using System ...
- python模块(os)
os模块 os模块提供了许多与操作系统交互的接口 os.getcwd() -> str # 返回当前路径, 相当于pwd os.chdir("dirname") -> ...
- Git 使用的想法
git rebase 每个提交(commit)取消掉,并且把它们临时 保存为补丁(patch)(这些补丁放到".git/rebase"目录中),然后 git fetch origi ...
- snapshot standby database
快照备库接收和归档主库发送来的redo,但是不会应用:切换成physical standby之后会自动开启redo apply.快照standby不可以参加主备切换:在最大保护性模式下,如果只有一个备 ...
- mysql explain输出中type的取值说明
原文: http://www.cnitblog.com/aliyiyi08/archive/2008/09/09/48878.html 这列很重要,显示了连接使用了哪种连接类别,有无使用索引. 从最好 ...
- Nand Flash与Nor Flash的区别
区别:http://zhidao.baidu.com/question/1068445.html?qbl=relate_question_0&word=Serial%20Flash%20%D3 ...
- csu oj 1330 字符识别？
http://acm.csu.edu.cn/OnlineJudge/problem.php?id=1330 1330: 字符识别? Time Limit: 1 Sec Memory Limit: 1 ...
- C#控件：TabControl
tabcontrol在切换页面的时候经常用到 这里讲下如何让tabcontrol更好看 ref:http://blog.csdn.net/conmajia/article/details/759671 ...
- Maven2的配置文件settings.xml(转)
当Maven运行过程中的各种配置,例如pom.xml,不想绑定到一个固定的project或者要分配给用户时,我们使用settings.xml中的settings元素来确定这些配置.这包含了本地仓库位置 ...