A deep dive into serialization and deserialization in PHP-PHP Tutorial-php.cn

This article will give you an in-depth analysis of serialization and deserialization in PHP. It has certain reference value. Friends in need can refer to it. I hope it will be helpful to everyone.

A deep dive into serialization and deserialization in PHP

##[Recommended learning: "

PHP Video Tutorial"]

Serialization

Serialization format

In PHP, serialization is used to store or transfer PHP values without losing their type and structure.

The serialization function prototype is as follows:

string serialize ( mixed $value )

Copy after login

First look at the following example:

class CC {
	public $data;
	private $pass;

	public function __construct($data, $pass)
	{
		$this->data = $data;
		$this->pass = $pass;
	}
}
$number = 34;
$str = 'uusama';
$bool = true;
$null = NULL;
$arr = array('a' => 1, 'b' => 2);
$cc = new CC('uu', true);

var_dump(serialize($number));
var_dump(serialize($str));
var_dump(serialize($bool));
var_dump(serialize($null));
var_dump(serialize($arr));
var_dump(serialize($cc));

Copy after login

The output result is:

string(5) "i:34;"
string(13) "s:6:"uusama";"
string(4) "b:1;"
string(2) "N;"
string(30) "a:2:{s:1:"a";i:1;s:1:"b";i:2;}"
string(52) "O:2:"CC":2:{s:4:"data";s:2:"uu";s:8:" CC pass";b:1;}"

Copy after login

So serialization for different types gets The string format is:

String : s:size:value;
Integer : i:value;
Boolean : b:value;(save 1 or 0)
Null : N;
Array : a:size:{key definition;value definition;(repeated per element)}
Object : O:strlen(object name):object name:object size:{s:strlen (property name):property name:property definition;(repeated per property)}

Serialized object

As we can see from the above example When serializing an object, only the property values are saved.

If it is inheritance, will the variables of the parent class be saved?

class CB {
	public $CB_data = 'cb';
}

class CC extends CB{
	const SECOND = 60;

	public $data;
	private $pass;

	public function __construct($data, $pass)
	{
		$this->data = $data;
		$this->pass = $pass;
	}

	public function setPass($pass)
	{
		$this->pass = $pass;
	}
}
$cc = new CC('uu', true);

var_dump(serialize($cc));

Copy after login

The output result is:

string(75) "O:2:"CC":3:{s:4:"data";s:2:"uu";s:8:" CC pass";b:1;s:7:"CB_data";s:2:"cb";}"

Copy after login

Obviously, when the object is serialized, it will not be saved. The value of the constant. Variables in the parent class will be retained.

Object serialization customization

When serializing an object, we do not need to save some sensitive attributes in the object. How should we deal with this?

When calling the

serialize() function to serialize an object, the function checks whether there is a magic method __sleep() in the class. If present, this method will be called first, and then the serialization operation will be performed. You can customize the serialization behavior by overloading this method. The prototype of this method is as follows:

public array __sleep ( void )

Copy after login

This method does not return anything, then NULL is serialized , and generates an
E_NOTICE level error
__sleep() cannot return the name of the private member of the parent class. Doing so will generate an E_NOTICE level error. At this time, the Serializable interface can only be used instead.

Look at the following example:

class User{
	const SITE = 'uusama';

	public $username;
	public $nickname;
	private $password;

	public function __construct($username, $nickname, $password)
	{
		$this->username = $username;
		$this->nickname = $nickname;
		$this->password = $password;
	}

	// 重载序列化调用的方法
	public function __sleep()
	{
		// 返回需要序列化的变量名，过滤掉password变量
		return array('username', 'nickname');
	}
}
$user = new User('uusama', 'uu', '123456');
var_dump(serialize($user));

Copy after login

The return result is as follows, obviously The value of the password field is ignored during serialization.

string(67) "O:4:"User":2:{s:8:"username";s:6:"uusama";s:8:"nickname";s:2:"uu";}"

Copy after login

Serialized object storage

Through the above introduction, we can serialize a copied object or data into a sequence string, and colleagues who save the value also save their structure.

We can save the serialized value and store it in a file or cache. It is not recommended to store it in the database because of the readability problem, and it is not easy to migrate and maintain, and it is not easy to query.

$user = new User('uusama', 'uu', '123456');
$ser = serialize($user);
// 保存在本地
file_put_contents('user.ser', $ser);

Copy after login

Deserialization

Usage method

Through the above explanation, we can serialize the object into a string And save it, so how to restore these serialized strings to their original state? PHP provides the deserialization function:

mixed unserialize ( string $str )

Copy after login

unserialize()The deserialization function is used to convert a single serialized variable back to a PHP value.

E_NOTICE is generated.
integer``float, string, array or object
__wakeup() member function (if it exists)

See the following example:

class User{
	const SITE = 'uusama';

	public $username;
	public $nickname;
	private $password;
	private $order;

	public function __construct($username, $nickname, $password)
	{
		$this->username = $username;
		$this->nickname = $nickname;
		$this->password = $password;
	}

	// 定义反序列化后调用的方法
	public function __wakeup()
	{
		$this->password = $this->username;
	}
}
$user_ser = 'O:4:"User":2:{s:8:"username";s:6:"uusama";s:8:"nickname";s:2:"uu";}';
var_dump(unserialize($user_ser));

Copy after login

The output result is:

object(User)#1 (4) {
  ["username"]=>
  string(6) "uusama"
  ["nickname"]=>
  string(2) "uu"
  ["password":"User":private]=>
  string(6) "uusama"
  ["order":"User":private]=>
  NULL
}

Copy after login

The following conclusion can be drawn:

The function is executed after the object is constructed, so $this->username The value is not emptyWhen deserializing, the variable value will be matched and copied to the serialized object as much as possible

Processing of undefined classes

In the above example, we defined the

User

class in advance before calling the deserialization function unserialize(). What would happen if we did not define it? Woolen cloth? <div class="code" style="position:relative; padding:0px; margin:0px;"><pre class="brush:php;toolbar:false">$user_ser = 'O:4:"User":2:{s:8:"username";s:6:"uusama";s:8:"nickname";s:2:"uu";}'; var_dump(unserialize($user_ser));</pre><div class="contentsignin">Copy after login</div></div>In this example, we did not define any

User

class. The deserialization was executed normally and no error was reported. The result obtained is as follows: <div class="code" style="position:relative; padding:0px; margin:0px;"><pre class="brush:php;toolbar:false">object(__PHP_Incomplete_Class)#1 (3) { ["__PHP_Incomplete_Class_Name"]=> string(4) "User" ["username"]=> string(6) "uusama" ["nickname"]=> string(2) "uu" }</pre><div class="contentsignin">Copy after login</div></div>Note before comparison The result of defining the

User

class, the object obtained by deserialization here is __PHP_Incomplete_Class, and the class name of the undefined class is specified. <p>如果这个时候我们去使用这个反序列化后的不明对象，则会抛出<code>E_NOTICE。这么看着不能用也不是办法，那么如何处理呢？有两种方案。

定义__autoload()等函数，指定发现未定义类时加载类的定义文件
可通过 php.ini、ini_set() 或 .htaccess 定义unserialize_callback_func。每次实例化一个未定义类时它都会被调用

以上两种方案的实现如下：

// unserialize_callback_func 从 PHP 4.2.0 起可用
ini_set('unserialize_callback_func', 'mycallback'); // 设置您的回调函数
function mycallback($classname) 
{
   // 只需包含含有类定义的文件
   // $classname 指出需要的是哪一个类
}


// 建议使用下面的函数，代替__autoload()
spl_autoload_register(function ($class_name) {
	// 动态加载未定义类的定义文件
    require_once $class_name . '.php';
});

Copy after login

PHP预定义序列化接口`Serializable`

还记得上面在将序列化过程中遇到的：无法在__sleep()方法中返回父类对象的问题吗，方法就是实现序列化接口Serializable。

该接口的原型如下：

Serializable {
	abstract public string serialize ( void )
	abstract public mixed unserialize ( string $serialized )
}

Copy after login

需要注意的是，如果定义的类实现了Serializable接口，那么序列化和反序列化的时候，PHP就不会再去调用__sleep()方法和__wakeup()方法。

class CB implements Serializable{
	public $CB_data = '';
	private $CB_password = 'ttt';

	public function setCBPassword($password)
	{
		$this->CB_password = $password;
	}

	public function serialize()
	{
		echo __METHOD__ . "\n";
		return serialize($this->CB_password);
	}

	public function unserialize($serialized)
	{
		echo __METHOD__ . "\n";
	}
}

class CC extends CB {
	const SECOND = 60;

	public $data;
	private $pass;

	public function __construct($data, $pass)
	{
		$this->data = $data;
		$this->pass = $pass;
	}

	public function __sleep()
	{
		// 输出调用了该方法名
		echo __METHOD__ . "\n";
	}

	public function __wakeup()
	{
		// 输出调用了该方法名
		echo __METHOD__ . "\n";
	}
}
$cc = new CC('uu', true);
$ser = serialize($cc);
var_dump($ser);
$un_cc = unserialize($ser);
var_dump($un_cc);

Copy after login

运行结果为：

CB::serialize
string(24) "C:2:"CC":10:{s:3:"ttt";}"
CB::unserialize
object(CC)#2 (4) {
  ["data"]=>
  NULL
  ["pass":"CC":private]=>
  NULL
  ["CB_data"]=>
  string(0) ""
  ["CB_password":"CB":private]=>
  string(3) "ttt"
}

Copy after login

可以完全定义serialize()方法，该方法返回的值就是序列化后大括号内的值，只要保证自定义序列化和反序列化的规则一致即可。

题外话

在PHP应用中，序列化和反序列化一般用做缓存，比如session缓存，cookie等。

序列化和反序列化在PHP中用得不算多，在Java语言中用得比较多。其实你有没有发现，这种把一个对象或者数组的变量转化成字符串的方式，json也可以做到。

使用json来实现对象和字符串之间的转换，在PHP中显得更加直观和轻便。而且经过测试，使用json_encode()比serialize()方法更加快速，大概快2~3倍。

在我看来，序列化和反序列化是一种传输抽象数据的思想。通过定义序列化和反序列化的规则，我们可以实现将PHP中的对象序列化成字节流，然后传输给别的语言或者系统使用，这在远程调用里面非常的方便。

更多编程相关知识，请访问：编程视频！！

The above is the detailed content of A deep dive into serialization and deserialization in PHP. For more information, please follow other related articles on the PHP Chinese website!