This project is no longer maintained and has been archived.


Behaviors


Introduction

Many times you may find classes having similar things within your models. These things may contain anything related to the schema of the component itself (relations, column definitions, index definitions etc.). One obvious way of re-factoring the code is having a base class with some classes extending it.

However inheritance solves only a fraction of things. The following sections show how using Doctrine_Template is much more powerful and flexible than using inheritance.

Doctrine_Template is a class template system. Templates are basically ready-to-use little components that your Record classes can load. When a template is being loaded its setTableDefinition() and setUp() methods are being invoked and the method calls inside them are being directed into the class in question.

This chapter describes the usage of various behaviors available for Doctrine. You'll also learn how to create your own behaviors. In order to grasp the concepts of this chapter you should be familiar with the theory behind Doctrine_Template and Doctrine_Record_Generator. We will explain what these classes are shortly.

When referring to behaviors we refer to class packages that use templates, generators and listeners extensively. All the introduced components in this chapter can be considered core behaviors, that means they reside at the Doctrine main repository.

Usually behaviors use generators side-to-side with template classes (classes that extend Doctrine_Template). The common workflow is:

  • A new template is being initialized
  • The template creates the generator and calls initialize() method
  • The template is attached to given class

As you may already know templates are used for adding common definitions and options to record classes. The purpose of generators is much more complex. Usually they are being used for creating generic record classes dynamically. The definitions of these generic classes usually depend on the owner class. For example the columns of the AuditLog versioning class are the columns of the parent class with all the sequence and autoincrement definitions removed.


Simple Templates

In the following example we define a template called TimestampBehavior. Basically the purpose of this template is to add date columns 'created' and 'updated' to the record class that loads this template. Additionally this template uses a listener called Timestamp listener which updates these fields based on record actions.

// models/TimestampListener.php
class TimestampListener extends Doctrine_Record_Listener
{
    public function preInsert(Doctrine_Event $event)
    {
        $event->getInvoker()->created = date('Y-m-d', time());
        $event->getInvoker()->updated = date('Y-m-d', time());
    }

    public function preUpdate(Doctrine_Event $event)
    {
        $event->getInvoker()->updated = date('Y-m-d', time());
    }
}

Now lets create a child Doctrine_Template named TimestampTemplate so we can attach it to our models with the actAs() method:

$ // models/TimestampBehavior.php

class TimestampTemplate extends Doctrine_Template { public function setTableDefinition() { $this->hasColumn('created', 'date'); $this->hasColumn('updated', 'date');

$this->addListener(new TimestampListener()); } }

Lets say we have a class called BlogPost that needs the timestamp functionality. All we need to do is to add actAs() call in the class definition.

class BlogPost extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('title', 'string', 200);
        $this->hasColumn('body', 'clob');
    }

    public function setUp()
    {
        $this->actAs('TimestampBehavior');
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- BlogPost: actAs: [TimestampBehavior] columns: title: string(200) body: clob
2
3
4
5
6

Now when we try and utilize the BlogPost model you will notice that the created and updated columns were added for you and automatically set when saved:

$ blogPost        = new BlogPost();
$ blogPost->title = 'Test';
$ blogPost->body  = 'test';
$ blogPost->save();

print_r($blogPost->toArray());

The above example would produce the following output:

$ php test.php
Array
(
    [id] => 1
    [title] => Test
    [body] => test
    [created] => 2009-01-22
    [updated] => 2009-01-22
)

The above described functionality is available via the Timestampable behavior that we have already talked about. You can go back and read more about it in the section of this chapter.


Templates with Relations

Many times the situations tend to be much more complex than the situation in the previous chapter. You may have model classes with relations to other model classes and you may want to replace given class with some extended class.

Consider we have two classes, User and Email, with the following definitions:

class User extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('username', 'string', 255);
        $this->hasColumn('password', 'string', 255);
    }

    public function setUp()
    {
        $this->hasMany('Email', array(
                'local' => 'id',
                'foreign' => 'user_id'
            )
        );
    }
}

class Email extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('address', 'string');
        $this->hasColumn('user_id', 'integer');
    }

    public function setUp()
    {
        $this->hasOne('User', array(
                'local' => 'user_id',
                'foreign' => 'id'
            )
        );
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- User: columns: username: string(255) password: string(255) Email: columns: address: string user_id: integer relations: User:
2
3
4
5
6
7
8
9
10
11
12

Now if we extend the User and Email classes and create, for example, classes ExtendedUser and ExtendedEmail, the ExtendedUser will still have a relation to the Email class and not the ExtendedEmail class. We could of course override the setUp() method of the User class and define relation to the ExtendedEmail class, but then we lose the whole point of inheritance. Doctrine_Template can solve this problem elegantly with its dependency injection solution.

In the following example we'll define two templates, UserTemplate and EmailTemplate, with almost identical definitions as the User and Email class had.

// models/UserTemplate.php
class UserTemplate extends Doctrine_Template
{
    public function setTableDefinition()
    {
        $this->hasColumn('username', 'string', 255);
        $this->hasColumn('password', 'string', 255);
    }

    public function setUp()
    {
        $this->hasMany('EmailTemplate as Emails', array(
                'local' => 'id',
                'foreign' => 'user_id'
            )
        );
    }
}

Now lets define the EmailTemplate:

// models/EmailTemplate.php
class EmailTemplate extends Doctrine_Template
{
    public function setTableDefinition()
    {
        $this->hasColumn('address', 'string');
        $this->hasColumn('user_id', 'integer');
    }

    public function setUp()
    {
        $this->hasOne('UserTemplate as User', array(
                'local' => 'user_id',
                'foreign' => 'id'
            )
        );
    }
}

Notice how we set the relations. We are not pointing to concrete Record classes, rather we are setting the relations to templates. This tells Doctrine that it should try to find concrete Record classes for those templates. If Doctrine can't find these concrete implementations the relation parser will throw an exception, but before we go ahead of things, here are the actual record classes:

class User extends Doctrine_Record
{
    public function setUp()
    {
        $this->actAs('UserTemplate');
    }
}

class Email extends Doctrine_Record
{
    public function setUp()
    {
        $this->actAs('EmailTemplate');
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- User: actAs: [UserTemplate] Email: actAs: [EmailTemplate]
2
3
4
5
6

Now consider the following code snippet. This does NOT work since we haven't yet set any concrete implementations for the templates.

// test.php

// ...
$ user = new User();
$ user->Emails; // throws an exception

The following version works. Notice how we set the concrete implementations for the templates globally using Doctrine_Manager:

// bootstrap.php

// ...
$ manager->setImpl('UserTemplate', 'User')
        ->setImpl('EmailTemplate', 'Email');

Now this code will work and won't throw an exception like it did before:

$ user                     = new User();
$ user->Emails[0]->address = '[email protected]';
$ user->save();

print_r($user->toArray(true));

The above example would produce the following output:

$ php test.php
Array
(
    [id] => 1
    [username] =>
    [password] =>
    [Emails] => Array
        (
            [0] => Array
                (
                    [id] => 1
                    [address] => [email protected]
                    [user_id] => 1
                )
        )
)

The implementations for the templates can be set at manager, connection and even at the table level.


Delegate Methods

Besides from acting as a full table definition delegate system, Doctrine_Template allows the delegation of method calls. This means that every method within the loaded templates is available in the record that loaded the templates. Internally the implementation uses magic method called __call() to achieve this functionality.

Lets add to our previous example and add some custom methods to the UserTemplate:

// models/UserTemplate.php
class UserTemplate extends Doctrine_Template
{
    // ...
    public function authenticate($username, $password)
    {
        $invoker = $this->getInvoker();
        if ($invoker->username == $username && $invoker->password == $password)
        {
            return true;
        }
        else
        {
            return false;
        }
    }
}

Now take a look at the following code and how we can use it:

$ user           = new User();
$ user->username = 'jwage';
$ user->password = 'changeme';

if ($user->authenticate('jwage', 'changemte'))
{
    echo 'Authenticated successfully!';
}
else
{
    echo 'Could not authenticate user!';
}

You can also delegate methods to Doctrine_Table classes just as easily. But, to avoid naming collisions the methods for table classes must have the string TableProxy appended to the end of the method name.

Here is an example where we add a new finder method:

// models/UserTemplate.php
class UserTemplate extends Doctrine_Template
{
    // ...
    public function findUsersWithEmailTableProxy()
    {
        return Doctrine_Query::create()
            ->select('u.username')
            ->from('User u')
            ->innerJoin('u.Emails e')
            ->execute();
    }
}

Now we can access that function from the Doctrine_Table object for the User model:

$ userTable = Doctrine_Core::getTable('User');
$ users = $userTable->findUsersWithEmail();

Each class can consists of multiple templates. If the templates contain similar definitions the most recently loaded template always overrides the former.


Creating Behaviors

This subchapter provides you the means for creating your own behaviors. Lets say we have various different Record classes that need to have one-to-many emails. We achieve this functionality by creating a generic behavior which creates Email classes on the fly.

We start this task by creating a behavior called EmailBehavior with a setTableDefinition() method. Inside the setTableDefinition() method various helper methods can be used for easily creating the dynamic record definition. Commonly the following methods are being used:

public function initOptions()
public function buildLocalRelation()
public function buildForeignKeys(Doctrine_Table $table)
public function buildForeignRelation($alias = null)
public function buildRelation() // calls buildForeignRelation() and buildLocalRelation()
class EmailBehavior extends Doctrine_Record_Generator
{
    public function initOptions()
    {
        $this->setOption('className', '%CLASS%Email');

        // Some other options
        // $this->setOption('appLevelDelete', true);
        // $this->setOption('cascadeDelete', false);
    }

    public function buildRelation()
    {
        $this->buildForeignRelation('Emails');
        $this->buildLocalRelation();
    }

    public function setTableDefinition()
    {
        $this->hasColumn('address', 'string', 255, array(
                'email'  => true,
                'primary' => true
            )
        );
    }
}

Core Behaviors

For the next several examples using the core behaviors lets delete all our existing schemas and models from our test environment we created and have been using in the earlier chapters:

$ rm schema.yml
$ touch schema.yml
$ rm -rf models/*

Introduction

Doctrine comes bundled with some templates that offer out of the box functionality for your models. You can enable these templates in your models very easily. You can do it directly in your Doctrine_Records or you can specify them in your YAML schema if you are managing your models with YAML.

In the next several examples we will demonstrate some of the behaviors that come bundled with Doctrine.


Versionable

Lets create a BlogPost model that we want to have the ability to have versions:

// models/BlogPost.php
class BlogPost extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('title', 'string', 255);
        $this->hasColumn('body', 'clob');
    }

    public function setUp()
    {
        $this->actAs('Versionable', array(
                'versionColumn' => 'version',
                'className' => '%CLASS%Version',
                'auditLog' => true
            )
        );
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- BlogPost: actAs: Versionable: versionColumn: version className: %CLASS%Version auditLog: true columns: title: string(255) body: clob
2
3
4
5
6
7
8
9
10

The auditLog option can be used to turn off the audit log history. This is when you want to maintain a version number but not maintain the data at each version.

Lets check the SQL that is generated by the above models:

// test.php

// ...
$ sql = Doctrine_Core::generateSqlFromArray(array('BlogPost'));
echo $sql[0] . "\n";
echo $sql[1];

The above code would output the following SQL query:

CREATE TABLE blog_post_version (id BIGINT,
title VARCHAR(255),
body LONGTEXT,
version BIGINT,
PRIMARY KEY(id,
version)) ENGINE = INNODB
CREATE TABLE blog_post (id BIGINT AUTO_INCREMENT,
title VARCHAR(255),
body LONGTEXT,
version BIGINT,
PRIMARY KEY(id)) ENGINE = INNODB
ALTER TABLE blog_post_version ADD FOREIGN KEY (id) REFERENCES blog_post(id) ON UPDATE CASCADE ON DELETE CASCADE

Notice how we have 2 additional statements we probably didn't expect to see. The behavior automatically created a blog_post_version table and related it to blog_post.

Now when we insert or update a BlogPost the version table will store all the old versions of the record and allow you to revert back at anytime. When you instantiate a BlogPost for the first time this is what is happening internally:

  • It creates a class called BlogPostVersion on-the-fly, the table this record is pointing at is blog_post_version
  • Everytime a BlogPost object is deleted / updated the previous version is stored into blog_post_version
  • Everytime a BlogPost object is updated its version number is increased.

Now lets play around with the BlogPost model:

$ blogPost        = new BlogPost();
$ blogPost->title = 'Test blog post';
$ blogPost->body  = 'test';
$ blogPost->save();

$ blogPost->title = 'Modified blog post title';
$ blogPost->save();

print_r($blogPost->toArray());

The above example would produce the following output:

$ php test.php
Array
(
    [id] => 1
    [title] => Modified blog post title
    [body] => test
    [version] => 2
)

Notice how the value of the version column is 2. This is because we have saved 2 versions of the BlogPost model. We can easily revert to another version by using the revert() method that the behavior includes.

Lets revert back to the first version:

$ blogPost->revert(1);
print_r($blogPost->toArray());

The above example would produce the following output:

$ php test.php
Array
(
    [id] => 2
    [title] => Test blog post
    [body] => test
    [version] => 1
)

Notice how the value of the version column is set to 1 and the title is back to the original value was set it to when creating the BlogPost.


Timestampable

The Timestampable behavior will automatically add a created_at and updated_at column and automatically set the values when a record is inserted and updated.

Since it is common to want to know the date a post is made lets expand our BlogPost model and add the Timestampable behavior to automatically set these dates for us.

// models/BlogPost.php
class BlogPost extends Doctrine_Record
{
    // ...
    public function setUp()
    {
        $this->actAs('Timestampable');
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- # schema.yml # ... BlogPost: actAs: # ... Timestampable: # ...
2
3
4
5
6
7
8
9

If you are only interested in using only one of the columns, such as a created_at timestamp, but not a an updated_at field, set the disabled to true for either of the fields as in the example below.

1--- BlogPost: actAs: # ... Timestampable: created: name: created_at type: timestamp format: Y-m-d H:i:s updated: disabled: true # ...
2
3
4
5
6
7
8
9
10
11
12

Now look what happens when we create a new post:

$ blogPost        = new BlogPost();
$ blogPost->title = 'Test blog post';
$ blogPost->body  = 'test';
$ blogPost->save();

print_r($blogPost->toArray());

The above example would produce the following output:

$ php test.php
Array
(
    [id] => 1
    [title] => Test blog post
    [body] => test
    [version] => 1
    [created_at] => 2009-01-21 17:54:23
    [updated_at] => 2009-01-21 17:54:23
)

Look how the created_at and updated_at values were automatically set for you!

Here is a list of all the options you can use with the Timestampable behavior on the created side of the behavior:

Name Default Description
name created_at The name of the column.
type timestamp The column type.
options array() Any additional options for the column.
format Y-m-d H:i:s The format of the timestamp if you don't use the timestamp column type. The date is built using PHP's date() function.
disabled false Whether or not to disable the created date.
expression NOW() Expression to use to set the column value.

Here is a list of all the options you can use with the Timestampable behavior on the updated side of the behavior that are not possible on the created side:

Name Default Description
onInsert true Whether or not to set the updated date when the record is first inserted.

Sluggable

The Sluggable behavior is a nice piece of functionality that will automatically add a column to your model for storing a unique human readable identifier that can be created from columns like title, subject, etc. These values can be used for search engine friendly urls.

Lets expand our BlogPost model to use the Sluggable behavior because we will want to have nice URLs for our posts:

// models/BlogPost.php
class BlogPost extends Doctrine_Record
{
    // ...
    public function setUp()
    {
        // ...
        $this->actAs('Sluggable', array(
                'unique'    => true,
                'fields'    => array('title'),
                'canUpdate' => true
            )
        );
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- # schema.yml # ... BlogPost: actAs: # ... Sluggable: unique: true fields: [title] canUpdate: true # ...
2
3
4
5
6
7
8
9
10
11
12

Now look what happens when we create a new post. The slug column will automatically be set for us:

$ blogPost        = new BlogPost();
$ blogPost->title = 'Test blog post';
$ blogPost->body  = 'test';
$ blogPost->save();

print_r($blogPost->toArray());

The above example would produce the following output:

$ php test.php
Array
(
    [id] => 1
    [title] => Test blog post
    [body] => test
    [version] => 1
    [created_at] => 2009-01-21 17:57:05
    [updated_at] => 2009-01-21 17:57:05
    [slug] => test-blog-post
)

Notice how the value of the slug column was automatically set based on the value of the title column. When a slug is created, by default it is urlized which means all non-url-friendly characters are removed and white space is replaced with hyphens(-).

The unique flag will enforce that the slug created is unique. If it is not unique an auto incremented integer will be appended to the slug before saving to database.

The canUpdate flag will allow the users to manually set the slug value to be used when building the url friendly slug.

Here is a list of all the options you can use on the Sluggable behavior:

Name Default Description
name slug The name of the slug column.
alias null The alias of the slug column.
type string The type of the slug column.
length 255 The length of the slug column.
unique true Whether or not unique slug values are enforced.
options array() Any other options for the slug column.
fields array() The fields that are used to build slug value.
uniqueBy array() The fields that make determine a unique slug.
uniqueIndex true Whether or not to create a unique index.
canUpdate false Whether or not the slug can be updated.
builder array('Doctrine_Inflector', 'urlize') The Class::method() used to build the slug.
indexName sluggable The name of the index to create.

I18n

Doctrine_I18n package is a behavior for Doctrine that provides internationalization support for record classes. In the following example we have a NewsItem class with two fields title and content. We want to have the field title with different languages support. This can be achieved as follows:

class NewsItem extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('title', 'string', 255);
        $this->hasColumn('body', 'blog');
    }

    public function setUp()
    {
        $this->actAs('I18n', array(
                'fields' => array('title', 'body')
            )
        );
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- NewsItem: actAs: I18n: fields: [title, body] columns: title: string(255) body: clob
2
3
4
5
6
7
8

Below is a list of all the options you can use with the I18n behavior:

Name Default Description
className %CLASS%Translation The name pattern to use for generated class.
fields array() The fields to internationalize.
type string The type of lang column.
length 2 The length of the lang column.
options array() Other options for the lang column.

Lets check the SQL that is generated by the above models:

// test.php

// ...
$ sql = Doctrine_Core::generateSqlFromArray(array('NewsItem'));
echo $sql[0] . "";
echo $sql[1];

The above code would output the following SQL query:

CREATE TABLE news_item_translation (id BIGINT,
title VARCHAR(255),
body LONGTEXT,
lang CHAR(2),
PRIMARY KEY(id,
lang)) ENGINE = INNODB
CREATE TABLE news_item (id BIGINT AUTO_INCREMENT,
PRIMARY KEY(id)) ENGINE = INNODB

Notice how the field title is not present in the news_item table. Since its present in the translation table it would be a waste of resources to have that same field in the main table. Basically Doctrine always automatically removes all translated fields from the main table.

Now the first time you initialize a new NewsItem record Doctrine initializes the behavior that builds the followings things:

  1. Record class called NewsItemTranslation
  2. Bi-directional relations between NewsItemTranslation and NewsItem

Lets take a look at how we can manipulate the translations of the NewsItem:

// test.php

// ...
$ newsItem = new NewsItem();
$ newsItem->Translation['en']->title = 'some title';
$ newsItem->Translation['en']->body  = 'test';
$ newsItem->Translation['fi']->title = 'joku otsikko';
$ newsItem->Translation['fi']->body  = 'test'; $newsItem->save();

print_r($newsItem->toArray());

The above example would produce the following output:

$ php test.php
Array
(
    [id] => 1
    [Translation] => Array
        (
            [en] => Array
                (
                    [id] => 1
                    [title] => some title
                    [body] => test
                    [lang] => en
                )
            [fi] => Array
                (
                    [id] => 1
                    [title] => joku otsikko
                    [body] => test
                    [lang] => fi
                )
        )
)

How do we retrieve the translated data now? This is easy! Lets find all items and their Finnish translations:

// test.php

// ...
$ newsItems = Doctrine_Query::create()
    ->from('NewsItem n')
    ->leftJoin('n.Translation t')
    ->where('t.lang = ?')
    ->execute(array('fi'));

echo $newsItems[0]->Translation['fi']->title;

The above example would produce the following output:

$ php test.php
joku otsikko

NestedSet

The NestedSet behavior allows you to turn your models in to a nested set tree structure where the entire tree structure can be retrieved in one efficient query. It also provided a nice interface for manipulating the data in your trees.

Lets take a Category model for example where the categories need to be organized in a hierarchical tree structure:

// models/Category.php
class Category extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('name', 'string', 255);
    }

    public function setUp()
    {
        $this->actAs('NestedSet', array(
                'hasManyRoots' => true,
                'rootColumnName' => 'root_id'
            )
        );
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- # schema.yml # ... Category: actAs: NestedSet: hasManyRoots: true rootColumnName: root_id columns: name: string(255)
2
3
4
5
6
7
8
9
10
11

Lets check the SQL that is generated by the above models:

// test.php

// ...
$ sql = Doctrine_Core::generateSqlFromArray(array('Category'));
echo $sql[0];

The above code would output the following SQL query:

CREATE TABLE category (id BIGINT AUTO_INCREMENT,
name VARCHAR(255),
root_id INT,
lft INT,
rgt INT,
level SMALLINT,
PRIMARY KEY(id)) ENGINE = INNODB

Notice how the root_id, lft, rgt and level columns are automatically added. These columns are used to organize the tree structure and are handled automatically for you internally.

We won't discuss the NestedSet behavior in 100% detail here. It is a very large behavior so it has its own Hierarchical Data.


Searchable

The Searchable behavior is a fulltext indexing and searching tool. It can be used for indexing and searching both database and files.

Imagine we have a Job model for job postings and we want it to be easily searchable:

// models/Job.php
class Job extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('title', 'string', 255);
        $this->hasColumn('description', 'clob');
    }

    public function setUp()
    {
        $this->actAs('Searchable', array(
                'fields' => array('title', 'content')
            )
        );
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- Job: actAs: Searchable: fields: [title, description] columns: title: string(255) description: clob
2
3
4
5
6
7
8

Lets check the SQL that is generated by the above models:

// test.php

// ...
$ sql = Doctrine_Core::generateSqlFromArray(array('Job'));
echo $sql[0] . "";
echo $sql[1] . "";
echo $sql[2];

The above code would output the following SQL query:

CREATE TABLE job_index (id BIGINT,
keyword VARCHAR(200),
field VARCHAR(50),
position BIGINT,
PRIMARY KEY(id,
keyword,
field,
position)) ENGINE = INNODB
CREATE TABLE job (id BIGINT AUTO_INCREMENT,
title VARCHAR(255),
description LONGTEXT,
PRIMARY KEY(id)) ENGINE = INNODB
ALTER TABLE job_index ADD FOREIGN KEY (id) REFERENCES job(id) ON UPDATE CASCADE ON DELETE CASCADE

Notice how the job_index table is automatically created for you and a foreign key between job and job_index was automatically created.

Because the Searchable behavior is such a large topic, we have more information on this that can be found in the Searching chapter.


Geographical

The below is only a demo. The Geographical behavior can be used with any data record for determining the number of miles or kilometers between 2 records.

// models/Zipcode.php
class Zipcode extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('zipcode', 'string', 255);
        $this->hasColumn('city', 'string', 255);
        $this->hasColumn('state', 'string', 2);
        $this->hasColumn('county', 'string', 255);
        $this->hasColumn('zip_class', 'string', 255);
    }

    public function setUp()
    {
        $this->actAs('Geographical');
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- # schema.yml # ... Zipcode: actAs: [Geographical] columns: zipcode: string(255) city: string(255) state: string(2) county: string(255) zip_class: string(255)
2
3
4
5
6
7
8
9
10
11
12

Lets check the SQL that is generated by the above models:

// test.php

// ...
$ sql = Doctrine_Core::generateSqlFromArray(array('Zipcode'));
echo $sql[0];

The above code would output the following SQL query:

CREATE TABLE zipcode (id BIGINT AUTO_INCREMENT,
zipcode VARCHAR(255),
city VARCHAR(255),
state VARCHAR(2),
county VARCHAR(255),
zip_class VARCHAR(255),
latitude DOUBLE,
longitude DOUBLE,
PRIMARY KEY(id)) ENGINE = INNODB

Notice how the Geographical behavior automatically adds the latitude and longitude columns to the records used for calculating distance between two records. Below you will find some example usage.

First lets retrieve two different zipcode records:

// test.php

// ...
$ zipcode1 = Doctrine_Core::getTable('Zipcode')->findOneByZipcode('37209');
$ zipcode2 = Doctrine_Core::getTable('Zipcode')->findOneByZipcode('37388');

Now we can get the distance between those two records by using the getDistance() method that the behavior provides:

// test.php

// ...
echo $zipcode1->getDistance($zipcode2, $kilometers = false);

The 2nd argument of the getDistance() method is whether or not to return the distance in kilometers. The default is false.

Now lets get the 50 closest zipcodes that are not in the same city:

// test.php

// ...
$ q = $zipcode1->getDistanceQuery();

$ q->orderby('miles asc')
    ->addWhere($q->getRootAlias() . '.city != ?', $zipcode1->city)
    ->limit(50);

echo $q->getSqlQuery();

The above call to getSql() would output the following SQL query:

SELECT
z.id AS z**id,
z.zipcode AS z**zipcode,
z.city AS z**city,
z.state AS z**state,
z.county AS z**county,
z.zip_class AS z**zip_class,
z.latitude AS z**latitude,
z.longitude AS z**longitude,
((ACOS(SIN(* PI() / 180) * SIN(z.latitude * PI() / 180) + COS(* PI() / 180) * COS(z.latitude * PI() / 180) * COS((- z.longitude) * PI() / 180)) * 180 / PI()) * 60 * 1.1515) AS z**0,
((ACOS(SIN(* PI() / 180) * SIN(z.latitude * PI() / 180) + COS(* PI() / 180) * COS(z.latitude * PI() / 180) * COS((- z.longitude) * PI() / 180)) * 180 / PI()) * 60 * 1.1515 * 1.609344) AS z**1
FROM zipcode z
WHERE z.city != ?
ORDER BY z__0 asc
LIMIT 50

Notice how the above SQL query includes a bunch of SQL that we did not write. This was automatically added by the behavior to calculate the number of miles between records.

Now we can execute the query and use the calculated number of miles values:

// test.php

// ...
$ result = $q->execute();

foreach ($result as $zipcode) {
    echo $zipcode->city . " - " . $zipcode->miles . "";
    // You could also access $zipcode->kilometers
}

Get some sample zip code data to test this

http://www.populardata.com/zip_codes.zip

Download and import the csv file with the following function:

// test.php

// ...
function parseCsvFile($file, $columnheadings = false, $delimiter = ',', $enclosure = "\"")
{
    $row    = 1;
    $rows   = array();
    $handle = fopen($file, 'r');

    while (($data = fgetcsv($handle, 1000, $delimiter, $enclosure)) !== FALSE) {

        if (!($columnheadings == false) && ($row == 1)) {
            $headingTexts = $data;
        } elseif (!($columnheadings == false)) {
            foreach ($data as $key => $value) {
                unset($data[$key]);
                $data[$headingTexts[$key]] = $value;
            }
            $rows[] = $data;
        } else {
            $rows[] = $data;
        }
        $row++;
    }

    fclose($handle);
    return $rows;
}

$ array = parseCsvFile('zipcodes.csv', false);

foreach ($array as $key => $value) {
    $zipcode = new Zipcode();
    $zipcode->fromArray($value);
    $zipcode->save();
}

SoftDelete

The SoftDelete behavior is a very simple yet highly desired model behavior which overrides the delete() functionality and adds a deleted_at column. When delete() is called, instead of deleting the record from the database, a delete_at date is set. Below is an example of how to create a model with the SoftDelete behavior being used.

// models/SoftDeleteTest.php
class SoftDeleteTest extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('name', 'string', null, array(
                'primary' => true
            )
        );
    }

    public function setUp()
    {
        $this->actAs('SoftDelete');
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- # schema.yml # ... SoftDeleteTest: actAs: [SoftDelete] columns: name: type: string(255) primary: true
2
3
4
5
6
7
8
9
10

Lets check the SQL that is generated by the above models:

// test.php

// ...
$ sql = Doctrine_Core::generateSqlFromArray(array('SoftDeleteTest'));
echo $sql[0];

The above code would output the following SQL query:

CREATE TABLE soft_delete_test (name VARCHAR(255),
deleted_at DATETIME DEFAULT NULL,
PRIMARY KEY(name)) ENGINE = INNODB

Now lets put the behavior in action.

You are required to enable DQL callbacks in order for all executed queries to have the dql callbacks executed on them. In the SoftDelete behavior they are used to filter the select statements to exclude all records where the deleted_at flag is set with an additional WHERE condition.

Enable DQL Callbacks

// bootstrap.php

// ...
$ manager->setAttribute(Doctrine_Core::ATTR_USE_DQL_CALLBACKS, true);

Now save a new record so we can test the SoftDelete functionality:

// test.php

// ...
$ record       = new SoftDeleteTest();
$ record->name = 'new record';
$ record->save();

Now when we call delete() the deleted_at flag will be set to true:

// test.php

// ...
$ record->delete();

print_r($record->toArray());

The above example would produce the following output:

$ php test.php
Array
(
    [name] => new record
    [deleted_at] => 2009-09-01 00:59:01
)

Also, when we select some data the query is modified for you:

// test.php

// ...
$ q = Doctrine_Query::create()
    ->from('SoftDeleteTest t');

echo $q->getSqlQuery();

The above call to getSql() would output the following SQL query:

SELECT
s.name AS s**name,
s.deleted_at AS s**deleted_at
FROM soft_delete_test s
WHERE (s.deleted_at IS NULL)

Notice how the where condition is automatically added to only return the records that have not been deleted.

Now if we execute the query:

// test.php

// ...
$ count = $q->count();
echo $count;

The above would be echo 0 because it would exclude the record saved above because the delete flag was set.


Nesting Behaviors

Below is an example of several behaviors to give a complete wiki database that is versionable, searchable, sluggable, and full I18n.

class Wiki extends Doctrine_Record
{
    public function setTableDefinition()
    {
        $this->hasColumn('title', 'string', 255);
        $this->hasColumn('content', 'string');
    }

    public function setUp()
    {
        $options  = array('fields' => array('title', 'content'));
        $auditLog = new Doctrine_Template_Versionable($options);
        $search   = new Doctrine_Template_Searchable($options);
        $slug     = new Doctrine_Template_Sluggable(array(
                'fields' => array('title')
            )
        );
        $i18n = new Doctrine_Template_I18n($options);

        $i18n->addChild($auditLog)
            ->addChild($search)
            ->addChild($slug);

        $this->actAs($i18n);

        $this->actAs('Timestampable');
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- WikiTest: actAs: I18n: fields: [title, content] actAs: Versionable: fields: [title, content] Searchable: fields: [title, content] Sluggable: fields: [title] columns: title: string(255) content: string
2
3
4
5
6
7
8
9
10
11
12
13
14
15

The above example of nesting behaviors is currently broken in Doctrine. We are working furiously to come up with a backwards compatible fix. We will announce when the fix is ready and update the documentation accordingly.


Generating Files

By default with behaviors the classes which are generated are evaluated at run-time and no files containing the classes are ever written to disk. This can be changed with a configuration option. Below is an example of how to configure the I18n behavior to generate the classes and write them to files instead of evaluating them at run-time.

class NewsArticle extends Doctrine_Record
{
    public function setTableDefinition() {
        $this->hasColumn('title', 'string', 255);
        $this->hasColumn('body', 'string', 255); $this->hasColumn('author', 'string', 255);
    }

    public function setUp()
    {
        $this->actAs('I18n', array(
                'fields'          => array('title', 'body'),
                'generateFiles'   => true,
                'generatePath'    => '/path/to/generate'
            )
        );
    }
}

Here is the same example in YAML format. You can read more about YAML in the YAML Schema Files chapter:

1--- NewsArticle: actAs: I18n: fields: [title, body] generateFiles: true generatePath: /path/to/generate columns: title: string(255) body: string(255) author: string(255)
2
3
4
5
6
7
8
9
10
11

Now the behavior will generate a file instead of generating the code and using eval() to evaluate it at runtime.


Querying Generated Classes

If you want to query the auto generated models you will need to make sure the model with the behavior attached is loaded and initialized. You can do this by using the static Doctrine_Core::initializeModels() method.

For example if you want to query the translation table for a BlogPost model you will need to run the following code:

Doctrine_Core::initializeModels(array('BlogPost'));

$ q = Doctrine_Query::create()
    ->from('BlogPostTranslation t')
    ->where('t.id = ? AND t.lang = ?', array(1, 'en'));

$ translations = $q->execute();

This is required because the behaviors are not instantiated until the model is instantiated for the first time. The above initializeModels() method instantiates the passed models and makes sure the information is properly loaded in to the array of loaded models.


Conclusion

By now we should know a lot about Doctrine behaviors. We should know how to write our own for our models as well as how to use all the great behaviors that come bundled with Doctrine.

Now we are ready to move on to discuss the Searching behavior in more detail in the Searching chapter. As it is a large topic we have devoted an entire chapter to it.