一切源於一個實驗,請看下面的例子:
表:
複製代碼 代碼如下:CREATE TABLE IF NOT EXISTS `foo` (
`a` int(10) unsigned NOT NULL AUTO_INCREMENT,
`b` int(10) unsigned NOT NULL,
`c` varchar(100) NOT NULL,
PRIMARY KEY (`a`),
KEY `bar` (`b`,`a`)
) ENGINE=InnoDB;
CREATE TABLE IF NOT EXISTS `foo2` (
`a` int(10) unsigned NOT NULL AUTO_INCREMENT,
`b` int(10) unsigned NOT NULL,
`c` varchar(100) NOT NULL,
PRIMARY KEY (`a`),
KEY `bar` (`b`,`a`)
) ENGINE=MyISAM;
我往兩個表中插入了30w的資料(插入的時候效能差別InnoDB比MyISAM慢)
複製代碼 代碼如下:<?php
$host = '192.168.100.166';
$dbName = 'test';
$user = 'root';
$password = '';
$db = mysql_connect($host, $user, $password) or die('DB connect failed');
mysql_select_db($dbName, $db);
echo '===================InnoDB=======================' . "\r\n";
$start = microtime(true);
mysql_query("SELECT SQL_NO_CACHE SQL_CALC_FOUND_ROWS * FROM foo WHERE b = 1 LIMIT 1000, 10");
$end = microtime(true);
echo $end - $start . "\r\n";
echo '===================MyISAM=======================' . "\r\n";
$start = microtime(true);
mysql_query("SELECT SQL_NO_CACHE SQL_CALC_FOUND_ROWS * FROM foo2 WHERE b = 1 LIMIT 1000, 10");
$end = microtime(true);
echo $end - $start . "\r\n";
返回結果:
一次查詢就會差別這麼多!!InnoDB和MyISAM,趕緊分析分析為什麼。
首先是使用explain來進行查看
確定兩邊都沒有使用index,第二個查詢查的rows,並且MyISAM的查詢rows還比InnoDB少這麼多,反而是查詢慢於InnoDB!!這Y的有點奇怪。
沒事,還有一個牛掰工具profile
具體使用可以參考:http://dev.mysql.com/doc/refman/5.0/en/show-profile.html
使用方法簡單來說:
複製代碼 代碼如下:Mysql > set profiling = 1;
Mysql>show profiles;
Mysql>show profile for query 1;
這個資料中就可以看到MyISAM的Sending data比InnoDB的Sending data費時太多了。查看mysql文檔
http://dev.mysql.com/doc/refman/5.0/en/general-thread-states.html
Sending data
The thread is reading and processing rows for a SELECT statement, and sending data to the client. Because operations occurring during this this state tend to perform large amounts of disk access (reads), it is often the longest-running state over the lifetime of a given query.
Sending data是去磁碟中讀取select的結果,然後將結果返回給用戶端。這個過程會有大量的IO操作。你可以使用show profile cpu for query XX;來進行查看,發現MyISAM的CPU_system比InnnoDB大很多。至此可以得出結論是MyISAM進行表查詢(區別僅僅使用索引就可以完成的查詢)比InnoDB慢。