MyISAM引擎和Heap引擎执行速度的性能测试

时间：2008-03-23 来源：sdccf

来源：赛迪网作者：30933

MyISAM引擎和Heap引擎执行速度的性能测试：

测试环境：

CPU: Intel Pentium4 2.66GHz Memory: 1GB Disk: 73GB/SCSI OS: FreeBSD 4.11 PHP: PHP 5.2.1 MySQL: MySQL 4.1.23b

前期工作：

my.cnf:

max_heap_table_size = 128M

建表：

use test; -- -- Store engine heap -- CREATE TABLE `tbl_heap` ( `id` int(11) NOT NULL auto_increment, `name` varchar(32) NOT NULL default '', `email` varchar(32) NOT NULL default '', `summary` varchar(255) default '', KEY `id` (`id`) ) ENGINE=HEAP DEFAULT CHARSET=gbk; -- -- Store engine myisam -- CREATE TABLE `tbl_isam` ( `id` int(11) NOT NULL auto_increment, `name` varchar(32) NOT NULL default '', `email` varchar(32) NOT NULL default '', `summary` varchar(255) default '', KEY `id` (`id`) ) ENGINE=MyISAM DEFAULT CHARSET=gbk;

插入数据：

说明：每次都是空表插入数据

[插入10000 Record] Heap engine insert 10000 record used time: 3.5008587837219 MyISAM engine insert 10000 record used time: 4.5881390571594 [50000 Record] Heap engine insert 50000 record used time: 19.895354986191 MyISAM engine insert 50000 record used time: 33.866044998169 [100000 Record] Heap engine insert 100000 record used time: 36.200875997543 MyISAM engine insert 100000 record used time: 68.34194111824 [200000 Record] Heap engine insert 200000 record used time: 68.00207901001 MyISAM engine insert 200000 record used time: 125.26263713837

查询数据：

表里分表有：200000条记录，两个表数据一致

[直接select，10000次，每次取100条记录] Heap engine select 10000 times, 100 record used time: 12.122506141663 MyISAM engine select 10000 times, 100 record used time: 19.512896060944 [直接select，1000次，每次取10000条记录] Heap engine select 1000 times, 10000 record used time: 111.54126811028 MyISAM engine select 1000 record used time: 116.79438710213 [增加where条件，1000次，每次取10000条记录] Heap engine select 1000 times, 10000 record used time: 111.52102303505 MyISAM engine select 1000 times, 10000 record used time: 117.68481087685 [where条件，10000次，每次从1000条起，取1000条记录] Heap engine select 10000 times, 1000 record used time: 124.28988695145 MyISAM engine select 10000 times, 1000 record used time: 139.82107305527 [where条件增加like，10000次，每次从1000条起，取1000条记录] Heap engine select 10000 times, 1000 record used time: 145.43780493736 MyISAM engine select 10000 times, 1000 record used time: 163.56296992302 [where条件增加索引，10000次，每次从1000条起，取1000条记录] -- 建立索引 (在SQLyob下执行) ALTER TABLE tbl_heap ADD INDEX idx_name (name); ALTER TABLE tbl_isam ADD INDEX idx_name (name); Heap engine alter table add index used time: 2.078 MyISAM engine alter table add index used time: 13.516 Heap engine select 10000 times, 1000 record used time: 153.48922395706 MyISAM engine select 10000 times, 1000 record used time: 239.86818814278 PS: 不合适的索引还不如不要。。。-_-# [SQLyog下进行count操作] Heap engine select count used time: 4.53 MyISAM engine select count used time: 3.28 [SQLyob下进行select操作] 速度都很快，都在1秒一下，不论是1000条记录，还是200000条记录，个人猜测可能不准确

更新操作:

[更新所有name=heiyeluren的记录为heiyeluren2，在SQLyog下执行] Heap engine update used time: 2.500 MyISAM engine update used time: 16.000

删除操作:

[删除所有name=heiyeluren2的记录，在SQLyog下执行] Heap engine delete used time: 51.172 MyISAM engine delte used time: 5.578

总结：

大家可以发现，Heap在插入、查询、更新操作上明显要比MyISAM快，但是删除操作稍稍有些慢，可能跟它在内存中的存储结构有关系，所以我们完全可以把Heap作为我们一个MyISAM表的一个备份，比如可以保存一些实时性要求比较高的数据，比如点击量、评论量、固定的用户信息等等，因为这些信息普遍就是插入和查询操作。

其实，Heap并没有比MyISAM快太多，这样说明我的MyISAM速度还是非常快的，所以一般情况下，MyISAM能够满足大部分的应用，如果数据太多的话，大家可以考虑把部分常用的数据保存到Heap表中，同时我们也可以结合Memcache等缓存工具来辅助缓存数据。

测试代码:

<?php /** * function lib */ function conn(){ $host = "localhost"; $user = "root"; $pass = ""; $db = "test"; $conn = mysql_connect($host, $user, $pass); if (!$conn || !is_resource($conn)){ die("Connect to mysql failed: ". mysql_error()); } mysql_select_db($db); return $conn; } function query($conn, $sql){ $res = mysql_query($sql, $conn); if (!$res){ echo mysql_error()." "; return false; } return true; } function get_data($conn, $sql){ $res = mysql_query($sql, $conn); if (!$res){ echo mysql_error()." "; return false; } $result = array(); while($row = mysql_fetch_array($res)){ $result[] = $row; } return $result; } function get_time() { list($usec, $sec) = explode(" ", microtime()); return ((float)$usec + (float)$sec); } /** * Insert test */ $conn = conn(); $count = 200000; //Heap engine $s1 = get_time(); for($i=0; $i<$count; $i++){ query($conn, "insert into tbl_heap set name = 'heiyeluren', email='[email protected]', summary='This message is summary'"); } $e1 = get_time(); echo "Heap engine insert $count record used time: ". ($e1-$s1) ." "; //MyISAM engine $s2 = get_time(); for($i=0; $i<$count; $i++){ query($conn, "insert into tbl_isam set name = 'heiyeluren', email='[email protected]', summary='This message is summary'"); } $e2 = get_time(); echo "MyISAM engine insert $count record used time: ". ($e2-$s2) ." "; /** * Select test */ $count = 1000; $records = 10000; $s1 = get_time(); for($i=0; $i<$count; $i++){ $res = get_data($conn, "select * from tbl_heap limit $records"); unset($res); } $e1 = get_time(); echo "Heap engine select $count record used time: ". ($e1-$s1) ." "; $s2 = get_time(); for($i=0; $i<$count; $i++){ $res = get_data($conn, "select * from tbl_isam limit $records"); unset($res); } $e2 = get_time(); echo "MyISAM engine select $count record used time: ". ($e2-$s2) ." "; ?>