模型元数据（Models Meta-Data）¶

为了加快开发速度 Phalcon\Mvc\Model 将帮助你从数据库中查询与模型有关字段和约束。为了实现这一点将会用 Phalcon\Mvc\Model\MetaData 管理和缓存表元数据。

有时在使用模型时需要获取和字段相关的属性。获得元数据的方法，如下所示:

<?php

$robot      = new Robots();

// Get Phalcon\Mvc\Model\Metadata instance
$metaData   = $robot->getModelsMetaData();

// Get robots fields names
$attributes = $metaData->getAttributes($robot);
print_r($attributes);

// Get robots fields data types
$dataTypes = $metaData->getDataTypes($robot);
print_r($dataTypes);

缓存元数据（Caching Meta-Data）¶

Once the application is in a production stage, it is not necessary to query the meta-data of the table from the database system each time you use the table. This could be done caching the meta-data using any of the following adapters:

Adapter	Description	API
Memory	这个是默认适配器，将元数据缓存在内存中。当请求完成时，元数据缓存将被释放，适合在应用程序开发阶段使用，每个请求都会重新获取字段信息 \| Phalcon\Mvc\Model\MetaData\Memory
Session	该适配器将元数据缓存在 `$_SESSION` 超全局变量中。该适配器只适合在拥有少量模型的时候使用，每次启动新的 Session 该适配器缓存就会失效。必须在 `session_start()` 启动 Session 之后使用	Phalcon\Mvc\Model\MetaData\Session
Apc	该适配器将使用 Alternative PHP Cache (APC) 存储元数据。可以设置 lifetime 选项。推荐在生成阶段使用该适配器	Phalcon\Mvc\Model\MetaData\Apc
XCache	This adapter uses XCache to store the table meta-data. You can specify the lifetime of the meta-data with options. This is the most recommended way to store meta-data when the application is in production stage	Phalcon\Mvc\Model\MetaData\Xcache
Files	This adapter uses plain files to store meta-data. By using this adapter the disk-reading is increased but the database access is reduced	Phalcon\Mvc\Model\MetaData\Files
Cache	使用缓存组件，将元数据存储在缓存组件对应的存储服务器中	Phalcon\Mvc\Model\MetaData\Cache

As other ORM’s dependencies, the metadata manager is requested from the services container:

<?php

use Phalcon\Mvc\Model\MetaData\Apc as ApcMetaData;

$di['modelsMetadata'] = function () {

    // Create a meta-data manager with APC
    $metaData = new ApcMetaData(
        array(
            "lifetime" => 86400,
            "prefix"   => "my-prefix"
        )
    );

    return $metaData;
};

自定义元数据缓存 KEY¶

<?php

class Robots extends Phalcon\Mvc\Model {

    public function getMetadataCachekey() {
        return 'Robots-robots';
    }
}

元数据策略（Meta-Data Strategies）¶

As mentioned above the default strategy to obtain the model’s meta-data is database introspection. In this strategy, the information schema is used to know the fields in a table, its primary key, nullable fields, data types, etc.

You can change the default meta-data introspection in the following way:

<?php

use Phalcon\Mvc\Model\MetaData\Apc as ApcMetaData;

$di['modelsMetadata'] = function () {

    // Instantiate a meta-data adapter
    $metaData = new ApcMetaData(
        array(
            "lifetime" => 86400,
            "prefix"   => "my-prefix"
        )
    );

    // Set a custom meta-data introspection strategy
    $metaData->setStrategy(new MyIntrospectionStrategy());

    return $metaData;
};

数据库内部策略（Database Introspection Strategy）¶

This strategy doesn’t require any customization and is implicitly used by all the meta-data adapters.

注解策略（Annotations Strategy）¶

This strategy makes use of annotations to describe the columns in a model:

<?php

use Phalcon\Mvc\Model;

class Robots extends Model
{
    /**
     * @Primary
     * @Identity
     * @Column(type="integer", nullable=false)
     */
    public $id;

    /**
     * @Column(type="string", length=70, nullable=false)
     */
    public $name;

    /**
     * @Column(type="string", length=32, nullable=false)
     */
    public $type;

    /**
     * @Column(type="integer", nullable=false)
     */
    public $year;
}

Annotations must be placed in properties that are mapped to columns in the mapped source. Properties without the @Column annotation are handled as simple class attributes.

The following annotations are supported:

Name	Description
Primary	Mark the field as part of the table’s primary key
Identity	The field is an auto_increment/serial column
Column	This marks an attribute as a mapped column

The annotation @Column supports the following parameters:

Name	Description
type	The column’s type (string, integer, decimal, boolean)
length	The column’s length if any
nullable	Set whether the column accepts null values or not

The annotations strategy could be set up this way:

<?php

use Phalcon\Mvc\Model\MetaData\Apc as ApcMetaData;
use Phalcon\Mvc\Model\MetaData\Strategy\Annotations as StrategyAnnotations;

$di['modelsMetadata'] = function () {

    // Instantiate a meta-data adapter
    $metaData = new ApcMetaData(
        array(
            "lifetime" => 86400,
            "prefix"   => "my-prefix"
        )
    );

    // Set a custom meta-data database introspection
    $metaData->setStrategy(new StrategyAnnotations());

    return $metaData;
};

自定义元数据（Manual Meta-Data）¶

Phalcon can obtain the metadata for each model automatically without the developer must set them manually using any of the introspection strategies presented above.

The developer also has the option of define the metadata manually. This strategy overrides any strategy set in the meta-data manager. New columns added/modified/removed to/from the mapped table must be added/modified/removed also for everything to work properly.

The following example shows how to define the meta-data manually:

<?php

use Phalcon\Mvc\Model;
use Phalcon\Db\Column;
use Phalcon\Mvc\Model\MetaData;

class Robots extends Model
{
    public function metaData()
    {
        return array(
            // Every column in the mapped table
            MetaData::MODELS_ATTRIBUTES => array(
                'id', 'name', 'type', 'year'
            ),

            // Every column part of the primary key
            MetaData::MODELS_PRIMARY_KEY => array(
                'id'
            ),

            // Every column that isn't part of the primary key
            MetaData::MODELS_NON_PRIMARY_KEY => array(
                'name', 'type', 'year'
            ),

            // Every column that doesn't allows null values
            MetaData::MODELS_NOT_NULL => array(
                'id', 'name', 'type'
            ),

            // Every column and their data types
            MetaData::MODELS_DATA_TYPES => array(
                'id'   => Column::TYPE_INTEGER,
                'name' => Column::TYPE_VARCHAR,
                'type' => Column::TYPE_VARCHAR,
                'year' => Column::TYPE_INTEGER
            ),

            // The columns that have numeric data types
            MetaData::MODELS_DATA_TYPES_NUMERIC => array(
                'id'   => true,
                'year' => true
            ),

            // The identity column, use boolean false if the model doesn't have
            // an identity column
            MetaData::MODELS_IDENTITY_COLUMN => 'id',

            // How every column must be bound/casted
            MetaData::MODELS_DATA_TYPES_BIND => array(
                'id'   => Column::BIND_PARAM_INT,
                'name' => Column::BIND_PARAM_STR,
                'type' => Column::BIND_PARAM_STR,
                'year' => Column::BIND_PARAM_INT
            ),

            // Fields that must be ignored from INSERT SQL statements
            MetaData::MODELS_AUTOMATIC_DEFAULT_INSERT => array(
                'year' => true
            ),

            // Fields that must be ignored from UPDATE SQL statements
            MetaData::MODELS_AUTOMATIC_DEFAULT_UPDATE => array(
                'year' => true
            ),

            // Default values for columns
            MetaData::MODELS_DEFAULT_VALUES => array(
                'year' => '2015'
            ),

            // Fields that allow empty strings
            MetaData::MODELS_EMPTY_STRING_VALUES => array(
                'name' => true
            )
        );
    }
}

Documentation