学步园

导读

上一篇文章提到count — Count all elements in an array, or something in an object，但是还有一个sizeof()，它是count的别名，为何会有一个别名出现呢?不会是写这个函数的人偏好C，而用了sizeof吧，决定一探究竟。

说明

以下测试将分作两个版本，一是一般数组大小 10000 个元素，二是大数组 100000 个元素，并且只采用cli方式执行。

详细

A 10000

<?php
echo 'Using for loop of sizeof(), the upper limit is 10,000. Do pre_calc or not, here are the performance differents >> ---- ';
$x = array_fill(0, 10000, 'a');	// 填充数组，数组长度为 1000， 值大小为 9.5367431640625MB

// 提前计算好大小
function with_pre_calc() {

	global $x;
	$lst = microtime(true);			// 循环开始时间 loop start time 浮点型
	$size = sizeof($x);
	for ($i=0; $i<$size; $i++);
	return microtime(true) - $lst;
}

// 没有提前计算好大小
function without_pre_calc() {

	global $x;
	$lst = microtime(true);			// 循环开始时间 loop start time 浮点型
	for ($i = 0; $i < sizeof($x); $i++);
	return microtime(true) - $lst;
}

echo 'With_pre_calc, using time >> ' . with_pre_calc() . ' second'
	. ' ---- ' 
	. 'Without_pre_calc, using time >> ' . without_pre_calc() . ' second'
	. ' ---- ';
	
unset($x);
echo 'Total execute the script, using time >> ' . (microtime(true) - $_SERVER['REQUEST_TIME_FLOAT']) . ' second';

<?php
echo 'Using for loop of count(), the upper limit is 10,000. Do pre_calc or not, here are the performance differents >> ---- ';
$x = array_fill(0, 10000, 'a');	// 填充数组，数组长度为 1000， 值大小为 9.5367431640625MB

// 提前计算好大小
function with_pre_calc() {

	global $x;
	$lst = microtime(true);			// 循环开始时间 loop start time 浮点型
	$size = count($x);
	for ($i=0; $i<$size; $i++);
	return microtime(true) - $lst;
}

// 没有提前计算好大小
function without_pre_calc() {

	global $x;
	$lst = microtime(true);			// 循环开始时间 loop start time 浮点型
	for ($i = 0; $i < count($x); $i++);
	return microtime(true) - $lst;
}

echo 'With_pre_calc, using time >> ' . with_pre_calc() . ' second'
	. ' ---- ' 
	. 'Without_pre_calc, using time >> ' . without_pre_calc() . ' second'
	. ' ---- ';
	
unset($x);
echo 'Total execute the script, using time >> ' . (microtime(true) - $_SERVER['REQUEST_TIME_FLOAT']) . ' second';

结果

Using for loop of sizeof(), the upper limit is 10,000. Do pre_calc or not, here are the performance differents >>

---- With_pre_calc, using time >> 0.0028781890869141 second 

---- Without_pre_calc, using time >> 11.052927017212 second 

---- Total execute the script, using time >> 11.059575080872 second

Using for loop of count(), the upper limit is 10,000. Do pre_calc or not, here are the performance differents >> 

---- With_pre_calc, using time >> 0.0029749870300293 second 

---- Without_pre_calc, using time >> 10.996887922287 second

---- Total execute the script, using time >> 11.003705024719 second

这似乎没有多大区别。

B 100000

（脚本中只是替换了数组大小）

结果（各运行十次，取均值）

Using for loop of sizeof(), the upper limit is 20,000. Do pre_calc or not, here are the performance differents >>

---- With_pre_calc, using time >> 0.0057840347290039 second 

---- Without_pre_calc, using time >> 54.608668088913 second 

---- Total execute the script, using time >> 54.619874000549 second

Using for loop of count(), the upper limit is 20,000. Do pre_calc or not, here are the performance differents >> 

---- With_pre_calc, using time >> 0.005860902557373 second 

---- Without_pre_calc, using time >> 55.447510004044 second 

---- Total execute the script, using time >> 55.458760023117 second

结论

这么看来，他们是没有区别的。难道真的是习惯问题？