system system.base system.caching system.caching.dependencies system.collections system.console system.db system.db.ar system.db.schema system.db.schema.mssql system.db.schema.mysql system.db.schema.oci system.db.schema.pgsql system.db.schema.sqlite system.i18n system.i18n.gettext system.logging system.utils system.validators system.web system.web.actions system.web.auth system.web.filters system.web.helpers system.web.renderers system.web.services system.web.widgets system.web.widgets.captcha system.web.widgets.pagers

CActiveRecord

system.db.ar
继承 abstract class CActiveRecord » CModel » CComponent
实现 ArrayAccess, Traversable, IteratorAggregate
可用自 1.0
版本 $Id$
CActiveRecord is the base class for classes representing relational data.

It implements the active record design pattern, a popular Object-Relational Mapping (ORM) technique.

Each active record class represents a database table/view, while an active record instance represents a row in that table. Using CActiveRecord, one can perform CRUD (create, read, update, delete) database actions in an object-oriented way. For example, the following code inserts a row to the 'Post' table (where 'Post' is the class representing the 'Post' table):
$post=new Post;
$post->title='sample post';
$post->save();


In the following, we elaborate how to use CActiveRecord.

First, we need to set up a database connection to be used by CActiveRecord. This is done by loading a CDbConnection application component whose ID is "db", using the following application configuration:
array(
    'components'=>array(
        'db'=>array(
            'class'=>'CDbConnection',
            'connectionString'=>'sqlite:test.db',
        ),
    )
);
To use a different database connection, you should override getDbConnection.

Second, for every database table that we want to access, define an active record class. The following example shows the minimal code needed for a 'Post' class, which represents the 'Post' table.
class Post extends CActiveRecord
{
    public static function model($className=__CLASS__)
    {
        return parent::model($className);
    }
}
The 'model()' method is declared as such for every active record class (to be explained shortly). By convention, the 'Post' class is associated with the database table named 'Post' (the class name). If the class should be associated with some other table, you may override the tableName() method.

To access column values, use $record->columnName, where $record refers to an active record instance. For example, the following code sets the 'title' column (attribute) of $post:
$post=new Post;
$post->title='a sample post';
Although we never explicitly declare the 'title' property in the 'Post' class, we can still access it in the above code. This is because 'title' is a column in the 'Post' table, and CActiveRecord makes it accessible as a property with the help of PHP __get() magic method. If you attempt to access a non-existing column in the same way, an exception will be thrown.

To insert a new row into 'Post', we first create a new instance of 'Post' class, and then call 'save()' to do the insertion.
$post=new Post;
$post->title='sample post';
$post->save();
After insertion, the $post object will contain an updated primary key if it is auto-incremental.

To query for posts, we will use the 'model' method we defined earlier on. The 'model' method is the only static method defined in CActiveRecord. It returns a static active record instance that is used to access class-level methods (something similar to static class methods). The following 'find' methods are all class-level methods that CActiveRecord implements to facilitate database querying:
// find the first row satisfying the specified condition
$post=Post::model()->find($condition,$params);

// find all rows satisfying the specified condition
$posts=Post::model()->findAll($condition,$params);

// find the row with the specified primary key
$post=Post::model()->findByPk($postID,$condition,$params);

// find all rows with the specified primary keys
$posts=Post::model()->findAllByPk($postIDs,$condition,$params);

// find the row with the specified attribute values
$post=Post::model()->findByAttributes($attributes,$condition,$params);

// find all rows with the specified attribute values
$posts=Post::model()->findAllByAttributes($attributes,$condition,$params);

// find the first row using the specified SQL statement
$post=Post::model()->findBySql($sql,$params);

// find all rows using the specified SQL statement
$posts=Post::model()->findAllBySql($sql,$params);

// get the number of rows satisfying the specified condition
$n=Post::model()->count($condition,$params);

// get the number of rows using the specified SQL statement
$n=Post::model()->countBySql($sql,$params);
where $condition specifies the WHERE clause, and $params gives a list of parameters that should be bound to the generated SQL statement. You may also pass a CDbCriteria object at the place of $condition to specify more complex query conditions (in that case, $params will be ignored since you can sepcify it in the CDbCriteria object.)

As we can see from the above, we have a set of find() methods and a set of findAll() methods. The result of the former is either an active record instance or null if no result is found, while the result of the latter is always an array.

After obtaining an active record from query, we can update it or delete it.
$post->title='new post title';
$post->save(); // or $post->delete();
In the above, we are using the same 'save()' method to do both insertion and update. CActiveRecord is intelligent enough to differentiate these two scenarios.

CActiveRecord also has a few class-level methods to facilitate updating and deleting rows without instantiating active record objects. Below is a summary:
// update all records with the specified attributes and condition
Post::model()->updateAll($attributes,$condition,$params);

// update one or several records with the specified primary key(s) and attribute values
Post::model()->updateByPk($postID,$attributes,$condition,$params);

// update one or several counter columns
Post::model()->updateCounters($counters,$condition,$params);

// delete one or several records with the specified primary key(s) and attribute values
Post::model()->deleteByPk($pk,$condition,$params);

// delete all records with the specified condition
Post::model()->deleteAll($condition,$params);

// delete all records which match the specified attribute values
Post::model()->deleteAllByAttributes($condition,$params);

// check if a record exists with the specified condition
Post::model()->exists($condition,$params);


A very useful feature that CActiveRecord supports is retrieving related active records. An active record is often related with other active records, via the relationship defined between database tables. For example, a post belongs to an author and has many comments; a user has a profile; a post belongs to and has many categories. CActiveRecord makes retrieving these related objects very easy.

Before retrieving related objects, we need to declare these relations in the class definition. This is done by overriding the relations method:
class Post extends CActiveRecord
{
    public function relations()
    {
        return array(
            'author'=>array(self::BELONGS_TO, 'User', 'authorID'),
            'comments'=>array(self::HAS_MANY, 'Comment', 'postID'),
        );
    }
}
In the above, we declare two relations:
  • the 'author' relation is of BELONGS_TO type; it references the 'User' active record class; and the 'authorID' column of the 'Post' table is the foreign key.
  • the 'comments' relation is of HAS_MANY type; it references the 'Comment' active record class; and the 'postID' column of the 'Comment' table is the foreign key.


Since the 'Post' class knows what kind of related objects it has, we can use the following code to access the author and comments of a post:
$post=Post::model()->findByPk($postID);
$author=$post->author;
$comments=$post->comments;
Internally, when we access 'author' or 'comments' the first time, an SQL query will be executed to fetch the corresponding active record(s). This is the so-called lazy loading, i.e., the related objects are loaded the first time when they are accessed.

If we have 100 posts and would like to obtain the author and comments for every post, lazy loading would be very inefficient, because it means we will need 200 SQL queries in order to fetch those related objects. In this situation, we should resort to the so-called eager loading.
$posts=Post::model()->with('author','comments')->findAll($condition,$params);
Here we used the 'with' method to specify that we want to bring back post records together with their related author and comments. Internally, two SQL queries will be created and executed: one brings back the posts and their authors, and the other brings back the comments). The reason we do not fetch them all together in one shot is because the returned rows would otherwise contain many repetitive results (e.g. the post information is repeated for every comment it has), which is both inefficient for database server and PHP code.

Eager loading can be nested. For example, if for every comment, we also want to know its author, we could use the following 'with' find:
$posts=Post::model()->with(array('author','comments'=>'author'))->findAll($condition,$params);
How many SQL queries do we need now? Still two (one for posts and their authors, and one for comments and their authors)! In fact, if there are N HAS_MANY/MANY_MANY relations involved, we would need N+1 queries in total.

There is a minor inconvenience when using eager loading, though. That is, we need to disambiguate any column references in conditions or search criteria, because we are joining several tables together. In general, it would be safe if we prefix every column referenced in the condition or search criteria with the table name.

Now let's describe more about the possible relations we can declare for active records. From database point of view, the relationship between two tables A and B has three types: one-to-many, one-to-one and many-to-many. In active records, these are classified into four types:
  • BELONGS_TO: if A vs. B is one-to-many, then B belongs to A (e.g. a post belongs to an author);
  • HAS_MANY: if A vs. B is one-to-many, then A has many B (e.g. an author has many posts);
  • HAS_ONE: this is special case of has-many where A has at most one B;
  • MANY_MANY: this corresponds to many-to-many relationship in database. An intermediate join table is needed as most DBMS do not support many-to-many relationship directly. For example, a post can belong to many categories and a category can have many posts. We would use a table 'post_category' to join the 'posts' and 'categories' tables together.


We already see how to declare a relation in an active record class. In many situations, we often have more constraints on those related objects, such as the conditions that the related objects should satisfy, their orderings, etc. These can be specified as additional options in our relation declaration. For example,
class Post extends CActiveRecord
{
    ......
    public function relations()
    {
        return array(
            'comments'=>array(self::HAS_MANY,'Comment','postID',
                              'order'=>'??.createTime DESC',
                              'with'=>'author'),
        );
    }
}
where the 'order' option specifies that the comments should be sorted by their creation time, and the 'with' option specifies that the comments should be loaded together with their authors. The special token '??.' is for disamibiguating the column reference. When SQL statement is generated, it will be replaced automatically by the table alias for the 'Comment' table. More detailed description about possible options can be found in CActiveRelation and CHasManyRelation.

Since version 1.0.4, a new relational query, called STAT, is supported. It is meant to perform relational queries that return statistical information about related objects (e.g. the number of comments that each post has). To use this new relational query, we need to declare it in the relations method like other relations:
class Post extends CActiveRecord
{
    ......
    public function relations()
    {
        return array(
            'commentCount'=>array(self::STAT,'Comment','postID'),

            'comments'=>array(self::HAS_MANY,'Comment','postID',
                              'order'=>'??.createTime DESC',
                              'with'=>'author'),
        );
    }
}
In the above, we declare 'commentCount' to be a STAT relation that is related with the Comment model via its foreign key postID. Given a post model instance, we can then obtain the number of comments it has via the expression $post->commentCount. We can also obtain the number of comments for a list of posts via eager loading:
$posts=Post::model()->with('commentCount')->findAll();




CActiveRecord has built-in validation functionality that validates the user input data before they are saved to database. To use the validation, override CModel::rules() as follows,
class Post extends CActiveRecord
{
    public function rules()
    {
        return array(
            array('title, content', 'required'),
            array('title', 'length', 'min'=>5),
        );
    }
}
The method returns a list of validation rules, each represented as an array of the following format:
array('attribute list', 'validator name', 'on'=>'insert', ...validation parameters...)
where
  • attribute list: specifies the attributes to be validated;
  • validator name: specifies the validator to be used. It can be a class name (or class in dot syntax), or a validation method in the AR class. A validator class must extend from CValidator, while a validation method must have the following signature:
    function validatorName($attribute,$params)
    
    When using a built-in validator class, you can use an alias name instead of the full class name. For example, you can use "required" instead of "system.validators.CRequiredValidator". For more details, see CValidator.
  • on: this specifies the scenario when the validation rule should be performed. Please see CModel::validate for more details about this option. NOTE: if the validation is triggered by the save call, it will use 'insert' to indicate an insertion operation, and 'update' an updating operation. You may thus specify a rule with the 'on' option so that it is applied only to insertion or updating.
  • additional parameters are used to initialize the corresponding validator properties. See CValidator for possible properties.


When save is called, validation will be performed. If there are any validation errors, save will return false and the errors can be retrieved back via getErrors.

公共属性

隐藏继承的属性

属性类型描述被定义在
attributes array Returns all column attribute values. CActiveRecord
commandBuilder CDbCommandBuilder the command builder used by this AR CActiveRecord
db CDbConnection the default database connection for all active record classes. CActiveRecord
dbConnection CDbConnection Returns the database connection used by active record. CActiveRecord
dbCriteria CDbCriteria Returns the query criteria associated with this model. CActiveRecord
errors array Returns the errors for all attribute or a single attribute. CModel
isNewRecord boolean whether the record is new and should be inserted when calling save. CActiveRecord
iterator CMapIterator Returns an iterator for traversing the attributes in the model. CModel
metaData CActiveRecordMetaData the meta for this AR class. CActiveRecord
primaryKey mixed the primary key value. CActiveRecord
safeAttributeNames array Returns the attribute names that are safe to be massively assigned. CModel
scenario string Returns the scenario that this model is in. CModel
tableSchema CDbTableSchema the metadata of the table that this AR belongs to CActiveRecord
validators array Returns a list of validators created according to rules. CActiveRecord

公共方法

隐藏继承的方法

方法描述被定义在
__call() Calls the named method which is not a class method. CActiveRecord
__construct() Constructor. CActiveRecord
__get() PHP getter magic method. CActiveRecord
__isset() Checks if a property value is null. CActiveRecord
__set() PHP setter magic method. CActiveRecord
__sleep() PHP sleep magic method. CActiveRecord
__unset() Sets a component property to be null. CActiveRecord
addError() Adds a new error to the specified attribute. CModel
addErrors() Adds a list of errors. CModel
addRelatedRecord() Adds a related object to this record. CActiveRecord
afterFindInternal() Calls afterFind. CActiveRecord
applyScopes() Applies the query scopes to the given criteria. CActiveRecord
asa() Returns the named behavior object. CComponent
attachBehavior() Attaches a behavior to this component. CComponent
attachBehaviors() Attaches a list of behaviors to the component. CComponent
attachEventHandler() Attaches an event handler to an event. CComponent
attributeLabels() Returns the attribute labels. CModel
attributeNames() Returns the list of all attribute names of the model. CActiveRecord
beforeFindInternal() Calls beforeFind. CActiveRecord
behaviors() Returns a list of behaviors that this model should behave as. CModel
canGetProperty() Determines whether a property can be read. CComponent
canSetProperty() Determines whether a property can be set. CComponent
clearErrors() Removes errors for all attributes or a single attribute. CModel
count() Finds the number of rows satisfying the specified query condition. CActiveRecord
countBySql() Finds the number of rows using the given SQL statement. CActiveRecord
createValidators() CModel
defaultScope() Returns the default named scope that should be implicitly applied to all queries for this model. CActiveRecord
delete() Deletes the row corresponding to this active record. CActiveRecord
deleteAll() Deletes rows with the specified condition. CActiveRecord
deleteAllByAttributes() Deletes rows which match the specified attribute values. CActiveRecord
deleteByPk() Deletes rows with the specified primary key. CActiveRecord
detachBehavior() Detaches a behavior from the component. CComponent
detachBehaviors() Detaches all behaviors from the component. CComponent
detachEventHandler() Detaches an existing event handler. CComponent
disableBehavior() Disables an attached behavior. CComponent
disableBehaviors() Disables all behaviors attached to this component. CComponent
enableBehavior() Enables an attached behavior. CComponent
enableBehaviors() Enables all behaviors attached to this component. CComponent
equals() Compares this active record with another one. CActiveRecord
exists() Checks whether there is row satisfying the specified condition. CActiveRecord
find() Finds a single active record with the specified condition. CActiveRecord
findAll() Finds all active records satisfying the specified condition. CActiveRecord
findAllByAttributes() Finds all active records that have the specified attribute values. CActiveRecord
findAllByPk() Finds all active records with the specified primary keys. CActiveRecord
findAllBySql() Finds all active records using the specified SQL statement. CActiveRecord
findByAttributes() Finds a single active record that has the specified attribute values. CActiveRecord
findByPk() Finds a single active record with the specified primary key. CActiveRecord
findBySql() Finds a single active record with the specified SQL statement. CActiveRecord
generateAttributeLabel() Generates a user friendly attribute label. CModel
getActiveRelation() CActiveRecord
getAttribute() Returns the named attribute value. CActiveRecord
getAttributeLabel() Returns the text label for the specified attribute. CModel
getAttributes() Returns all column attribute values. CActiveRecord
getCommandBuilder() CActiveRecord
getDbConnection() Returns the database connection used by active record. CActiveRecord
getDbCriteria() Returns the query criteria associated with this model. CActiveRecord
getError() Returns the first error of the specified attribute. CModel
getErrors() Returns the errors for all attribute or a single attribute. CModel
getEventHandlers() Returns the list of attached event handlers for an event. CComponent
getIsNewRecord() CActiveRecord
getIterator() Returns an iterator for traversing the attributes in the model. CModel
getMetaData() CActiveRecord
getPrimaryKey() CActiveRecord
getRelated() Returns the related record(s). CActiveRecord
getSafeAttributeNames() Returns the attribute names that are safe to be massively assigned. CModel
getScenario() Returns the scenario that this model is in. CModel
getTableSchema() CActiveRecord
getValidators() Returns a list of validators created according to rules. CActiveRecord
getValidatorsForAttribute() Returns the validators that are applied to the specified attribute under the specified scenario. CModel
hasAttribute() CActiveRecord
hasErrors() Returns a value indicating whether there is any validation error. CModel
hasEvent() Determines whether an event is defined. CComponent
hasEventHandler() Checks whether the named event has attached handlers. CComponent
hasProperty() Determines whether a property is defined. CComponent
hasRelated() Returns a value indicating whether the named related object(s) has been loaded. CActiveRecord
init() Initializes this model. CActiveRecord
insert() Inserts a row into the table based on this active record attributes. CActiveRecord
isAttributeRequired() Returns a value indicating whether the attribute is required. CModel
model() Returns the static model of the specified AR class. CActiveRecord
offsetExists() Returns whether there is an element at the specified offset. CActiveRecord
offsetGet() Returns the element at the specified offset. CModel
offsetSet() Sets the element at the specified offset. CModel
offsetUnset() Unsets the element at the specified offset. CModel
onAfterConstruct() This event is raised after the record instance is created by new operator. CActiveRecord
onAfterDelete() This event is raised after the record is deleted. CActiveRecord
onAfterFind() This event is raised after the record is instantiated by a find method. CActiveRecord
onAfterSave() This event is raised after the record is saved. CActiveRecord
onAfterValidate() This event is raised after the validation is performed. CModel
onBeforeDelete() This event is raised before the record is deleted. CActiveRecord
onBeforeFind() This event is raised before an AR finder performs a find call. CActiveRecord
onBeforeSave() This event is raised before the record is saved. CActiveRecord
onBeforeValidate() This event is raised before the validation is performed. CModel
populateRecord() Creates an active record with the given attributes. CActiveRecord
populateRecords() Creates a list of active records based on the input data. CActiveRecord
primaryKey() Returns the primary key of the associated database table. CActiveRecord
raiseEvent() Raises an event. CComponent
refresh() Repopulates this active record with the latest data. CActiveRecord
refreshMetaData() Refreshes the meta data for this AR class. CActiveRecord
relations() This method should be overridden to declare related objects. CActiveRecord
rules() Returns the validation rules for attributes. CModel
safeAttributes() Returns the name of attributes that are safe to be massively assigned. CActiveRecord
save() Saves the current record. CActiveRecord
saveAttributes() Saves a selected list of attributes. CActiveRecord
scopes() Returns the declaration of named scopes. CActiveRecord
setAttribute() Sets the named attribute value. CActiveRecord
setAttributes() Sets the attribute values in a massive way. CModel
setIsNewRecord() CActiveRecord
setScenario() CModel
tableName() Returns the name of the associated database table. CActiveRecord
update() Updates the row represented by this active record. CActiveRecord
updateAll() Updates records with the specified condition. CActiveRecord
updateByPk() Updates records with the specified primary key(s). CActiveRecord
updateCounters() Updates one or several counter columns. CActiveRecord
validate() Performs the validation. CModel
with() Specifies which related objects should be eagerly loaded. CActiveRecord

受保护的方法

隐藏继承的方法

方法描述被定义在
afterConstruct() This method is invoked after a record instance is created by new operator. CActiveRecord
afterDelete() This method is invoked after deleting a record. CActiveRecord
afterFind() This method is invoked after each record is instantiated by a find method. CActiveRecord
afterSave() This method is invoked after saving a record. CActiveRecord
afterValidate() This method is invoked after validation ends. CModel
beforeDelete() This method is invoked before deleting a record. CActiveRecord
beforeFind() This method is invoked before an AR finder executes a find call. CActiveRecord
beforeSave() This method is invoked before saving a record (after validation, if any). CActiveRecord
beforeValidate() This method is invoked before validation starts. CModel
instantiate() Creates an active record instance. CActiveRecord

事件

隐藏继承的事件

事件描述被定义在
onBeforeSave This event is raised before the record is saved. CActiveRecord
onAfterSave This event is raised after the record is saved. CActiveRecord
onBeforeDelete This event is raised before the record is deleted. CActiveRecord
onAfterDelete This event is raised after the record is deleted. CActiveRecord
onAfterConstruct This event is raised after the record instance is created by new operator. CActiveRecord
onBeforeFind This event is raised before an AR finder performs a find call. CActiveRecord
onAfterFind This event is raised after the record is instantiated by a find method. CActiveRecord
onBeforeValidate This event is raised before the validation is performed. CModel
onAfterValidate This event is raised after the validation is performed. CModel

属性详情

attributes 属性
public array getAttributes(mixed $names=true)
public void setAttributes(array $values, mixed $scenario='')

Returns all column attribute values. Note, related objects are not returned.

commandBuilder 属性 只读

the command builder used by this AR

db 属性
public static CDbConnection $db;

the default database connection for all active record classes. By default, this is the 'db' application component.

dbConnection 属性 只读

Returns the database connection used by active record. By default, the "db" application component is used as the database connection. You may override this method if you want to use a different database connection.

dbCriteria 属性 只读 (自版本 v1.0.5 可用)
public CDbCriteria getDbCriteria(boolean $createIfNull=true)

Returns the query criteria associated with this model.

isNewRecord 属性
public boolean getIsNewRecord()
public void setIsNewRecord(boolean $value)

whether the record is new and should be inserted when calling save. This property is automatically set in constructor and populateRecord. Defaults to false, but it will be set to true if the instance is created using the new operator.

metaData 属性 只读

the meta for this AR class.

primaryKey 属性 只读
public mixed getPrimaryKey()

the primary key value. An array (column name=>column value) is returned if the primary key is composite. If primary key is not defined, null will be returned.

tableSchema 属性 只读

the metadata of the table that this AR belongs to

validators 属性 只读 (自版本 v1.0.1 可用)
public array getValidators()

Returns a list of validators created according to rules. This overrides the parent implementation so that the validators are only created once for each type of AR.

方法详情

__call() 方法 (自版本 v1.0.5 可用)
public mixed __call(string $name, array $parameters)
$name string the method name
$parameters array method parameters
{return} mixed the method return value

Calls the named method which is not a class method. Do not call this method. This is a PHP magic method that we override to implement the named scope feature.

__construct() 方法
public void __construct(array $attributes=array ( ), string $scenario='')
$attributes array initial attributes (name => value). The attributes are subject to filtering via setAttributes. If this parameter is null, nothing will be done in the constructor (this is internally used by populateRecord).
$scenario string scenario name. See setAttributes for more details about this parameter. This parameter has been available since version 1.0.2. As of version 1.0.4, this parameter will be used to set the scenario property of the model.

Constructor.

参见

__get() 方法
public mixed __get(string $name)
$name string property name
{return} mixed property value

PHP getter magic method. This method is overridden so that AR attributes can be accessed like properties.

参见

__isset() 方法 (自版本 v1.0.1 可用)
public boolean __isset(string $name)
$name string the property name or the event name
{return} boolean whether the property value is null

Checks if a property value is null. This method overrides the parent implementation by checking if the named attribute is null or not.

__set() 方法
public void __set(string $name, mixed $value)
$name string property name
$value mixed property value

PHP setter magic method. This method is overridden so that AR attributes can be accessed like properties.

__sleep() 方法
public void __sleep()

PHP sleep magic method. This method ensures that the model meta data reference is set to null.

__unset() 方法 (自版本 v1.0.1 可用)
public void __unset(string $name)
$name string the property name or the event name

Sets a component property to be null. This method overrides the parent implementation by clearing the specified attribute value.

addRelatedRecord() 方法
public void addRelatedRecord(string $name, mixed $record, mixed $index)
$name string attribute name
$record mixed the related record
$index mixed the index value in the related object collection. If true, it means using zero-based integer index. If false, it means a HAS_ONE or BELONGS_TO object and no index is needed.

Adds a related object to this record. This method is used internally by CActiveFinder to populate related objects.

afterConstruct() 方法
protected void afterConstruct()

This method is invoked after a record instance is created by new operator. The default implementation raises the onAfterConstruct event. You may override this method to do postprocessing after record creation. Make sure you call the parent implementation so that the event is raised properly.

afterDelete() 方法
protected void afterDelete()

This method is invoked after deleting a record. The default implementation raises the onAfterDelete event. You may override this method to do postprocessing after the record is deleted. Make sure you call the parent implementation so that the event is raised properly.

afterFind() 方法
protected void afterFind()

This method is invoked after each record is instantiated by a find method. The default implementation raises the onAfterFind event. You may override this method to do postprocessing after each newly found record is instantiated. Make sure you call the parent implementation so that the event is raised properly.

afterFindInternal() 方法 (自版本 v1.0.3 可用)
public void afterFindInternal()

Calls afterFind. This method is internally used.

afterSave() 方法
protected void afterSave()

This method is invoked after saving a record. The default implementation raises the onAfterSave event. You may override this method to do postprocessing after record saving. Make sure you call the parent implementation so that the event is raised properly.

applyScopes() 方法 (自版本 v1.0.12 可用)
public void applyScopes(CDbCriteria $criteria)
$criteria CDbCriteria the query criteria. This parameter may be modified by merging dbCriteria.

Applies the query scopes to the given criteria. This method merges dbCriteria with the given criteria parameter. It then resets dbCriteria to be null.

attributeNames() 方法 (自版本 v1.0.1 可用)
public array attributeNames()
{return} array list of attribute names.

Returns the list of all attribute names of the model. This would return all column names of the table associated with this AR class.

beforeDelete() 方法
protected boolean beforeDelete()
{return} boolean whether the record should be deleted. Defaults to true.

This method is invoked before deleting a record. The default implementation raises the onBeforeDelete event. You may override this method to do any preparation work for record deletion. Make sure you call the parent implementation so that the event is raised properly.

beforeFind() 方法 (自版本 v1.0.9 可用)
protected void beforeFind()

This method is invoked before an AR finder executes a find call. The find calls include find, findAll, findByPk, findAllByPk, findByAttributes and findAllByAttributes. The default implementation raises the onBeforeFind event. If you override this method, make sure you call the parent implementation so that the event is raised properly.

beforeFindInternal() 方法 (自版本 v1.0.11 可用)
public void beforeFindInternal()

Calls beforeFind. This method is internally used.

beforeSave() 方法
protected boolean beforeSave()
{return} boolean whether the saving should be executed. Defaults to true.

This method is invoked before saving a record (after validation, if any). The default implementation raises the onBeforeSave event. You may override this method to do any preparation work for record saving. Use isNewRecord to determine whether the saving is for inserting or updating record. Make sure you call the parent implementation so that the event is raised properly.

count() 方法
public integer count(mixed $condition='', array $params=array ( ))
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} integer the number of rows satisfying the specified query condition.

Finds the number of rows satisfying the specified query condition. See find() for detailed explanation about $condition and $params.

countBySql() 方法
public integer countBySql(string $sql, array $params=array ( ))
$sql string the SQL statement
$params array parameters to be bound to the SQL statement
{return} integer the number of rows using the given SQL statement.

Finds the number of rows using the given SQL statement.

defaultScope() 方法 (自版本 v1.0.5 可用)
public array defaultScope()
{return} array the query criteria. This will be used as the parameter to the constructor of CDbCriteria.

Returns the default named scope that should be implicitly applied to all queries for this model. Note, default scope only applies to SELECT queries. It is ignored for INSERT, UPDATE and DELETE queries. The default implementation simply returns an empty array. You may override this method if the model needs to be queried with some default criteria (e.g. only active records should be returned).

delete() 方法
public boolean delete()
{return} boolean whether the deletion is successful.

Deletes the row corresponding to this active record.

deleteAll() 方法
public integer deleteAll(mixed $condition='', array $params=array ( ))
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} integer the number of rows deleted

Deletes rows with the specified condition. See find() for detailed explanation about $condition and $params.

deleteAllByAttributes() 方法 (自版本 v1.0.9 可用)
public CActiveRecord deleteAllByAttributes(array $attributes, mixed $condition='', array $params=array ( ))
$attributes array list of attribute values (indexed by attribute names) that the active records should match. Since version 1.0.8, an attribute value can be an array which will be used to generate an IN condition.
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} CActiveRecord the record found. Null if none is found.

Deletes rows which match the specified attribute values. See find() for detailed explanation about $condition and $params.

deleteByPk() 方法
public integer deleteByPk(mixed $pk, mixed $condition='', array $params=array ( ))
$pk mixed primary key value(s). Use array for multiple primary keys. For composite key, each key value must be an array (column name=>column value).
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} integer the number of rows deleted

Deletes rows with the specified primary key. See find() for detailed explanation about $condition and $params.

equals() 方法
public boolean equals($record)
$record
{return} boolean whether the two active records refer to the same row in the database table.

Compares this active record with another one. The comparison is made by comparing the primary key values of the two active records.

exists() 方法
public boolean exists(mixed $condition, array $params=array ( ))
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} boolean whether there is row satisfying the specified condition.

Checks whether there is row satisfying the specified condition. See find() for detailed explanation about $condition and $params.

find() 方法
public CActiveRecord find(mixed $condition='', array $params=array ( ))
$condition mixed query condition or criteria. If a string, it is treated as query condition (the WHERE clause); If an array, it is treated as the initial values for constructing a CDbCriteria object; Otherwise, it should be an instance of CDbCriteria.
$params array parameters to be bound to an SQL statement. This is only used when the first parameter is a string (query condition). In other cases, please use CDbCriteria::params to set parameters.
{return} CActiveRecord the record found. Null if no record is found.

Finds a single active record with the specified condition.

findAll() 方法
public array findAll(mixed $condition='', array $params=array ( ))
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} array list of active records satisfying the specified condition. An empty array is returned if none is found.

Finds all active records satisfying the specified condition. See find() for detailed explanation about $condition and $params.

findAllByAttributes() 方法
public array findAllByAttributes(array $attributes, mixed $condition='', array $params=array ( ))
$attributes array list of attribute values (indexed by attribute names) that the active records should match. Since version 1.0.8, an attribute value can be an array which will be used to generate an IN condition.
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} array the records found. An empty array is returned if none is found.

Finds all active records that have the specified attribute values. See find() for detailed explanation about $condition and $params.

findAllByPk() 方法
public array findAllByPk(mixed $pk, mixed $condition='', array $params=array ( ))
$pk mixed primary key value(s). Use array for multiple primary keys. For composite key, each key value must be an array (column name=>column value).
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} array the records found. An empty array is returned if none is found.

Finds all active records with the specified primary keys. See find() for detailed explanation about $condition and $params.

findAllBySql() 方法
public array findAllBySql(string $sql, array $params=array ( ))
$sql string the SQL statement
$params array parameters to be bound to the SQL statement
{return} array the records found. An empty array is returned if none is found.

Finds all active records using the specified SQL statement.

findByAttributes() 方法
public CActiveRecord findByAttributes(array $attributes, mixed $condition='', array $params=array ( ))
$attributes array list of attribute values (indexed by attribute names) that the active records should match. Since version 1.0.8, an attribute value can be an array which will be used to generate an IN condition.
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} CActiveRecord the record found. Null if none is found.

Finds a single active record that has the specified attribute values. See find() for detailed explanation about $condition and $params.

findByPk() 方法
public CActiveRecord findByPk(mixed $pk, mixed $condition='', array $params=array ( ))
$pk mixed primary key value(s). Use array for multiple primary keys. For composite key, each key value must be an array (column name=>column value).
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} CActiveRecord the record found. Null if none is found.

Finds a single active record with the specified primary key. See find() for detailed explanation about $condition and $params.

findBySql() 方法
public CActiveRecord findBySql(string $sql, array $params=array ( ))
$sql string the SQL statement
$params array parameters to be bound to the SQL statement
{return} CActiveRecord the record found. Null if none is found.

Finds a single active record with the specified SQL statement.

getActiveRelation() 方法
public CActiveRelation getActiveRelation(string $name)
$name string the relation name
{return} CActiveRelation the named relation declared for this AR class. Null if the relation does not exist.

getAttribute() 方法
public mixed getAttribute(string $name)
$name string the attribute name
{return} mixed the attribute value. Null if the attribute is not set or does not exist.

Returns the named attribute value. If this is a new record and the attribute is not set before, the default column value will be returned. If this record is the result of a query and the attribute is not loaded, null will be returned. You may also use $this->AttributeName to obtain the attribute value.

参见

getAttributes() 方法
public array getAttributes(mixed $names=true)
$names mixed names of attributes whose value needs to be returned. If this is true (default), then all attribute values will be returned, including those that are not loaded from DB (null will be returned for those attributes). If this is null, all attributes except those that are not loaded from DB will be returned.
{return} array attribute values indexed by attribute names.

Returns all column attribute values. Note, related objects are not returned.

getCommandBuilder() 方法
public CDbCommandBuilder getCommandBuilder()
{return} CDbCommandBuilder the command builder used by this AR

getDbConnection() 方法
public CDbConnection getDbConnection()
{return} CDbConnection the database connection used by active record.

Returns the database connection used by active record. By default, the "db" application component is used as the database connection. You may override this method if you want to use a different database connection.

getDbCriteria() 方法 (自版本 v1.0.5 可用)
public CDbCriteria getDbCriteria(boolean $createIfNull=true)
$createIfNull boolean whether to create a criteria instance if it does not exist. Defaults to true.
{return} CDbCriteria the query criteria that is associated with this model. This criteria is mainly used by named scope feature to accumulate different criteria specifications.

Returns the query criteria associated with this model.

getIsNewRecord() 方法
public boolean getIsNewRecord()
{return} boolean whether the record is new and should be inserted when calling save. This property is automatically set in constructor and populateRecord. Defaults to false, but it will be set to true if the instance is created using the new operator.

getMetaData() 方法
public CActiveRecordMetaData getMetaData()
{return} CActiveRecordMetaData the meta for this AR class.

getPrimaryKey() 方法
public mixed getPrimaryKey()
{return} mixed the primary key value. An array (column name=>column value) is returned if the primary key is composite. If primary key is not defined, null will be returned.

getRelated() 方法 (自版本 v1.0.2 可用)
public mixed getRelated(string $name, boolean $refresh=false, array $params=array ( ))
$name string the relation name (see relations)
$refresh boolean whether to reload the related objects from database. Defaults to false.
$params array additional parameters that customize the query conditions as specified in the relation declaration. This parameter has been available since version 1.0.5.
{return} mixed the related object(s).

Returns the related record(s). This method will return the related record(s) of the current record. If the relation is HAS_ONE or BELONGS_TO, it will return a single object or null if the object does not exist. If the relation is HAS_MANY or MANY_MANY, it will return an array of objects or an empty array.

getTableSchema() 方法
public CDbTableSchema getTableSchema()
{return} CDbTableSchema the metadata of the table that this AR belongs to

getValidators() 方法 (自版本 v1.0.1 可用)
public array getValidators()
{return} array a list of validators created according to CModel::rules.

Returns a list of validators created according to rules. This overrides the parent implementation so that the validators are only created once for each type of AR.

hasAttribute() 方法
public boolean hasAttribute(string $name)
$name string attribute name
{return} boolean whether this AR has the named attribute (table column).

hasRelated() 方法 (自版本 v1.0.3 可用)
public booolean hasRelated(string $name)
$name string the relation name
{return} booolean a value indicating whether the named related object(s) has been loaded.

Returns a value indicating whether the named related object(s) has been loaded.

init() 方法 (自版本 v1.0.8 可用)
public void init()

Initializes this model. This method is invoked in the constructor right after scenario is set. You may override this method to provide code that is needed to initialize the model (e.g. setting initial property values.)

insert() 方法
public boolean insert(array $attributes=NULL)
$attributes array list of attributes that need to be saved. Defaults to null, meaning all attributes that are loaded from DB will be saved.
{return} boolean whether the attributes are valid and the record is inserted successfully.

Inserts a row into the table based on this active record attributes. If the table's primary key is auto-incremental and is null before insertion, it will be populated with the actual value after insertion. Note, validation is not performed in this method. You may call validate to perform the validation.

instantiate() 方法 (自版本 v1.0.2 可用)
protected CActiveRecord instantiate(array $attributes)
$attributes array list of attribute values for the active records.
{return} CActiveRecord the active record

Creates an active record instance. This method is called by populateRecord and populateRecords. You may override this method if the instance being created depends the attributes that are to be populated to the record. For example, by creating a record based on the value of a column, you may implement the so-called single-table inheritance mapping.

model() 方法
public static CActiveRecord model(string $className='CActiveRecord')
$className string active record class name.
{return} CActiveRecord active record model instance.

Returns the static model of the specified AR class. The model returned is a static instance of the AR class. It is provided for invoking class-level methods (something similar to static class methods.)

EVERY derived AR class must override this method as follows,

public static function model($className=__CLASS__)
{
    return parent::model($className);
}

offsetExists() 方法 (自版本 v1.0.2 可用)
public boolean offsetExists(mixed $offset)
$offset mixed the offset to check on
{return} boolean

Returns whether there is an element at the specified offset. This method is required by the interface ArrayAccess.

onAfterConstruct() 方法 (自版本 v1.0.2 可用)
public void onAfterConstruct(CEvent $event)
$event CEvent the event parameter

This event is raised after the record instance is created by new operator.

onAfterDelete() 方法 (自版本 v1.0.2 可用)
public void onAfterDelete(CEvent $event)
$event CEvent the event parameter

This event is raised after the record is deleted.

onAfterFind() 方法 (自版本 v1.0.2 可用)
public void onAfterFind(CEvent $event)
$event CEvent the event parameter

This event is raised after the record is instantiated by a find method.

onAfterSave() 方法 (自版本 v1.0.2 可用)
public void onAfterSave(CEvent $event)
$event CEvent the event parameter

This event is raised after the record is saved.

onBeforeDelete() 方法 (自版本 v1.0.2 可用)
public void onBeforeDelete(CEvent $event)
$event CEvent the event parameter

This event is raised before the record is deleted.

onBeforeFind() 方法 (自版本 v1.0.9 可用)
public void onBeforeFind(CEvent $event)
$event CEvent the event parameter

This event is raised before an AR finder performs a find call.

参见

onBeforeSave() 方法 (自版本 v1.0.2 可用)
public void onBeforeSave(CEvent $event)
$event CEvent the event parameter

This event is raised before the record is saved.

populateRecord() 方法
public CActiveRecord populateRecord(array $attributes, boolean $callAfterFind=true)
$attributes array attribute values (column name=>column value)
$callAfterFind boolean whether to call afterFind after the record is populated. This parameter is added in version 1.0.3.
{return} CActiveRecord the newly created active record. The class of the object is the same as the model class. Null is returned if the input data is false.

Creates an active record with the given attributes. This method is internally used by the find methods.

populateRecords() 方法
public array populateRecords(array $data, boolean $callAfterFind=true)
$data array list of attribute values for the active records.
$callAfterFind boolean whether to call afterFind after each record is populated. This parameter is added in version 1.0.3.
{return} array list of active records.

Creates a list of active records based on the input data. This method is internally used by the find methods.

primaryKey() 方法 (自版本 v1.0.4 可用)
public mixed primaryKey()
{return} mixed the primary key of the associated database table. If the key is a single column, it should return the column name; If the key is a composite one consisting of several columns, it should return the array of the key column names.

Returns the primary key of the associated database table. This method is meant to be overridden in case when the table is not defined with a primary key (for some legency database). If the table is already defined with a primary key, you do not need to override this method. The default implementation simply returns null, meaning using the primary key defined in the database.

refresh() 方法
public boolean refresh()
{return} boolean whether the row still exists in the database. If true, the latest data will be populated to this active record.

Repopulates this active record with the latest data.

refreshMetaData() 方法 (自版本 v1.0.8 可用)
public void refreshMetaData()

Refreshes the meta data for this AR class. By calling this method, this AR class will regenerate the meta data needed. This is useful if the table schema has been changed and you want to use the latest available table schema. Make sure you have called CDbSchema::refresh before you call this method. Otherwise, old table schema data will still be used.

relations() 方法
public array relations()
{return} array list of related object declarations. Defaults to empty array.

This method should be overridden to declare related objects.

There are four types of relations that may exist between two active record objects:

  • BELONGS_TO: e.g. a member belongs to a team;
  • HAS_ONE: e.g. a member has at most one profile;
  • HAS_MANY: e.g. a team has many members;
  • MANY_MANY: e.g. a member has many skills and a skill belongs to a member.


By declaring these relations, CActiveRecord can bring back related objects in either lazy loading or eager loading approach, and you save the effort of writing complex JOIN SQL statements.

Each kind of related objects is defined in this method as an array with the following elements:
'varName'=>array('relationType', 'className', 'foreignKey', ...additional options)
where 'varName' refers to the name of the variable/property that the related object(s) can be accessed through; 'relationType' refers to the type of the relation, which can be one of the following four constants: self::BELONGS_TO, self::HAS_ONE, self::HAS_MANY and self::MANY_MANY; 'className' refers to the name of the active record class that the related object(s) is of; and 'foreignKey' states the foreign key that relates the two kinds of active record. Note, for composite foreign keys, they must be listed together, separating with space or comma; and for foreign keys used in MANY_MANY relation, the joining table must be declared as well (e.g. 'joinTable(fk1, fk2)').

Additional options may be specified as name-value pairs in the rest array elements:
  • 'select': string|array, a list of columns to be selected. Defaults to '*', meaning all columns. Column names should be disambiguated if they appear in an expression (e.g. COUNT(??.name) AS name_count).
  • 'condition': string, the WHERE clause. Defaults to empty. Note, column references need to be disambiguated with prefix '??.' (e.g. ??.age>20)
  • 'order': string, the ORDER BY clause. Defaults to empty. Note, column references need to be disambiguated with prefix '??.' (e.g. ??.age DESC)
  • 'with': string|array, a list of child related objects that should be loaded together with this object. Note, this is only honored by lazy loading, not eager loading.
  • 'joinType': type of join. Defaults to 'LEFT OUTER JOIN'.
  • 'aliasToken': the column prefix for column reference disambiguation. Defaults to '??.'.
  • 'alias': the alias for the table associated with this relationship. This option has been available since version 1.0.1. It defaults to null, meaning the table alias is automatically generated. This is different from `aliasToken` in that the latter is just a placeholder and will be replaced by the actual table alias.
  • 'params': the parameters to be bound to the generated SQL statement. This should be given as an array of name-value pairs. This option has been available since version 1.0.3.
  • 'on': the ON clause. The condition specified here will be appended to the joining condition using the AND operator. This option has been available since version 1.0.2.
  • 'index': the name of the column whose values should be used as keys of the array that stores related objects. This option is only available to HAS_MANY and MANY_MANY relations. This option has been available since version 1.0.7.


The following options are available for certain relations when lazy loading:
  • 'group': string, the GROUP BY clause. Defaults to empty. Note, column references need to be disambiguated with prefix '??.' (e.g. ??.age). This option only applies to HAS_MANY and MANY_MANY relations.
  • 'having': string, the HAVING clause. Defaults to empty. Note, column references need to be disambiguated with prefix '??.' (e.g. ??.age). This option only applies to HAS_MANY and MANY_MANY relations.
  • 'limit': limit of the rows to be selected. This option does not apply to BELONGS_TO relation.
  • 'offset': offset of the rows to be selected. This option does not apply to BELONGS_TO relation.


Below is an example declaring related objects for 'Post' active record class:
return array(
    'author'=>array(self::BELONGS_TO, 'User', 'authorID'),
    'comments'=>array(self::HAS_MANY, 'Comment', 'postID', 'with'=>'author', 'order'=>'createTime DESC'),
    'tags'=>array(self::MANY_MANY, 'Tag', 'PostTag(postID, tagID)', 'order'=>'name'),
);

safeAttributes() 方法
public array safeAttributes()
{return} array list of safe attribute names.

Returns the name of attributes that are safe to be massively assigned. The default implementation returns all table columns exception primary key(s). Child class may override this method to further limit the attributes that can be massively assigned. See CModel::safeAttributes on how to override this method.

save() 方法
public boolean save(boolean $runValidation=true, array $attributes=NULL)
$runValidation boolean whether to perform validation before saving the record. If the validation fails, the record will not be saved to database. The validation will be performed under either 'insert' or 'update' scenario, depending on whether isNewRecord is true or false.
$attributes array list of attributes that need to be saved. Defaults to null, meaning all attributes that are loaded from DB will be saved.
{return} boolean whether the saving succeeds

Saves the current record. The record is inserted as a row into the database table if it is manually created using the 'new' operator. If it is obtained using one of those 'find' methods, the record is considered not new and it will be used to update the corresponding row in the table. You may check this status via isNewRecord. Validation may be performed before saving the record. If the validation fails, the record will not be saved. If the record is being inserted and its primary key is null, after insertion the primary key will be populated with the value generated automatically by the database.

saveAttributes() 方法
public boolean saveAttributes(array $attributes)
$attributes array attributes to be updated. Each element represents an attribute name or an attribute value indexed by its name. If the latter, the record's attribute will be changed accordingly before saving.
{return} boolean whether the update is successful

Saves a selected list of attributes. Unlike save, this method only saves the specified attributes of an existing row dataset. It thus has better performance. Note, this method does neither attribute filtering nor validation. So do not use this method with untrusted data (such as user posted data). You may consider the following alternative if you want to do so:

$postRecord=Post::model()->findByPk($postID);
$postRecord->attributes=$_POST['post'];
$postRecord->save();

scopes() 方法 (自版本 v1.0.5 可用)
public array scopes()
{return} array the scope definition. The array keys are scope names; the array values are the corresponding scope definitions. Each scope definition is represented as an array whose keys must be properties of CDbCriteria.

Returns the declaration of named scopes. A named scope represents a query criteria that can be chained together with other named scopes and applied to a query. This method should be overridden by child classes to declare named scopes for the particular AR classes. For example, the following code declares two named scopes: 'recently' and 'published'.

return array(
    'published'=>array(
          'condition'=>'status=1',
    ),
    'recently'=>array(
          'order'=>'createTime DESC',
          'limit'=>5,
    ),
);
If the above scopes are declared in a 'Post' model, we can perform the following queries:
$posts=Post::model()->published()->findAll();
$posts=Post::model()->published()->recently()->findAll();
$posts=Post::model()->published()->with('comments')->findAll();
Note that the last query is a relational query.

setAttribute() 方法
public boolean setAttribute(string $name, mixed $value)
$name string the attribute name
$value mixed the attribute value.
{return} boolean whether the attribute exists and the assignment is conducted successfully

Sets the named attribute value. You may also use $this->AttributeName to set the attribute value.

参见

setIsNewRecord() 方法
public void setIsNewRecord(boolean $value)
$value boolean whether the record is new and should be inserted when calling save.

tableName() 方法
public string tableName()
{return} string the table name

Returns the name of the associated database table. By default this method returns the class name as the table name. You may override this method if the table is not named after this convention.

update() 方法
public boolean update(array $attributes=NULL)
$attributes array list of attributes that need to be saved. Defaults to null, meaning all attributes that are loaded from DB will be saved.
{return} boolean whether the update is successful

Updates the row represented by this active record. All loaded attributes will be saved to the database. Note, validation is not performed in this method. You may call validate to perform the validation.

updateAll() 方法
public integer updateAll(array $attributes, mixed $condition='', array $params=array ( ))
$attributes array list of attributes (name=>$value) to be updated
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} integer the number of rows being updated

Updates records with the specified condition. See find() for detailed explanation about $condition and $params. Note, the attributes are not checked for safety and no validation is done.

updateByPk() 方法
public integer updateByPk(mixed $pk, array $attributes, mixed $condition='', array $params=array ( ))
$pk mixed primary key value(s). Use array for multiple primary keys. For composite key, each key value must be an array (column name=>column value).
$attributes array list of attributes (name=>$value) to be updated
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} integer the number of rows being updated

Updates records with the specified primary key(s). See find() for detailed explanation about $condition and $params. Note, the attributes are not checked for safety and validation is NOT performed.

updateCounters() 方法
public integer updateCounters(array $counters, mixed $condition='', array $params=array ( ))
$counters array the counters to be updated (column name=>increment value)
$condition mixed query condition or criteria.
$params array parameters to be bound to an SQL statement.
{return} integer the number of rows being updated

Updates one or several counter columns. Note, this updates all rows of data unless a condition or criteria is specified. See find() for detailed explanation about $condition and $params.

with() 方法
public CActiveFinder with()
{return} CActiveFinder the active finder instance. If no parameter is passed in, the object itself will be returned.

Specifies which related objects should be eagerly loaded. This method takes variable number of parameters. Each parameter specifies the name of a relation or child-relation. For example,

// find all posts together with their author and comments
Post::model()->with('author','comments')->findAll();
// find all posts together with their author and the author's profile
Post::model()->with('author','author.profile')->findAll();
The relations should be declared in relations().

By default, the options specified in relations() will be used to do relational query. In order to customize the options on the fly, we should pass an array parameter to the with() method. The array keys are relation names, and the array values are the corresponding query options. For example,
Post::model()->with(array(
    'author'=>array('select'=>'id, name'),
    'comments'=>array('condition'=>'approved=1', 'order'=>'createTime'),
))->findAll();


This method returns a CActiveFinder instance that provides a set of find methods similar to that of CActiveRecord.

Note, the possible parameters to this method have been changed since version 1.0.2. Previously, it was not possible to specify on-th-fly query options, and child-relations were specified as hierarchical arrays.