Suppose we are operating a mobile phone-related website. Users can filter the mobile phones they want by specifying several parameters (such as operating system, screen resolution, camera pixels, etc.). However, because mobile phones have many parameters and the parameters of different mobile phones vary greatly, the parameter table structure is usually a vertical table (one parameter is one row) instead of a horizontal table (one parameter is one column). In this case, several parameters are used to obtain the results. , usually by taking each individual parameter to get the result, and then taking the intersection together. Assume that each parameter will contain about a thousand unique results (id int), and use this as a premise to simulate and generate some data:
Optimization of PHP array intersection
<font face="新宋体" size="2"><?php<br />
<br />
$rand = function() {<br />
$result = array();<br />
<br />
for ($i = 0; $i < 1000; null) {<br />
$value = mt_rand(1, 10000);<br />
<br />
if (!isset($result[$value])) {<br />
$result[$value] = null;<br />
$i++;<br />
}<br />
}<br />
<br />
return array_keys($result);<br />
};<br />
<br />
$param_a = $rand();<br />
$param_b = $rand();<br />
<br />
?></font>
<font face="新宋体" size="2"><?php<br />
<br />
$rand = function() {<br />
$result = array();<br />
<br />
for ($i = 0; $i < 1000; null) {<br />
$value = mt_rand(1, 10000);<br />
<br />
If (!isset($result[$value])) {<br />
$result[$value] = null;<br />
$i++;<br />
}<br />
}<br />
<br />
Return array_keys($result);<br />
};<br />
<br />
$param_a = $rand();<br />
$param_b = $rand();<br />
<br />
?></font>
Note: If the test data set is too small, the conclusions may be inconsistent. Let’s first take a look at the performance achieved through PHP’s built-in method array_intersect:
<font face="新宋体" size="2"><?php<br />
<br />
$time = microtime(true);<br />
<br />
$result = array_intersect($param_a, $param_b);<br />
<br />
$time = microtime(true) - $time;<br />
<br />
echo "array_intersect: {$time}n";<br />
<br />
?></font>
|
Let’s take a look at the performance achieved through the custom method intersect:
|
Intuitively, we would think that built-in functions are faster than custom functions, but in this case the opposite is true:
array_intersect: 0.023918151855469
intersect: 0.0026049613952637
We need to remind everyone that array_intersect and intersect are not completely equivalent in function. Examples are as follows:
<font face="新宋体" size="2">$param_a = array(1, 2, 2);<br>
$param_b = array(1, 2, 3);<br>
<br>
var_dump(<br>
Array_intersect($param_a, $param_b),<br>
Intersect($param_a, $param_b)<br>
);</font>
|
array_intersect: 1, 2, 2
intersect: 1, 2
In other words, if there are repeated elements in the first array parameter, array_intersect will return all repeated elements that meet the conditions, instead of just one. Interested readers can change the order of the parameters and see the results.
One more thing, when I first wrote the intersect method, it probably looked like this:
<font face="新宋体" size="2"><?php<br />
<br />
function intersect() {<br />
If (func_num_args() < 2) {<br />
trigger_error('param error', E_USER_ERROR);<br />
}<br />
<br />
$args = func_get_args();<br />
<br />
foreach ($args AS $arg) {<br />
if (!is_array($arg)) {<br />
trigger_error('param error', E_USER_ERROR);<br />
}<br />
}<br />
<br />
$result = array();<br />
<br />
$data = array_count_values(<br />
call_user_func_array('array_merge', $args)<br />
);<br />
<br />
foreach ($data AS $value => $count) {<br>
if ($count > 1) {<br>
$result[] = $value;<br>
}<br>
}<br>
<br>
Return $result;<br>
}<br>
<br>
?></font>
|
The code is more concise, but there is a drawback. Because array_merge is used, when there are very many elements in the array, the memory occupied will be relatively large. On the contrary, if there are not very many elements in the array, then this method is also feasible.
Reference: Faster array_intersect