Yii AR中处理多表关联的relations配置

关系型 Active Record官方文档中指出：
两张表之间的关联是根据外键来的，但是这种外键关联虽然在数据容错方面有益处，但是在性能上是个损伤，所以，一般是不定义外键的.
这种情况下，他们之间的关联又会根据什么来呢？
有A，B两张表，在A与B表中都存在一个字段filedX,并且同时字段filedX也是A,B的关联字段，这种情况下，可以直接
public function relations()
{
// NOTE: you may need to adjust the relation name and the related
// class name for the relations automatically generated below.
return array(
'VarName'=>array('RelationType', 'ClassName', 'filedX', ...additional options)
);
}

这样运用relations方法.

另外有一种情况是，虽然A，B两张表，有关联，但是关联字段不是同一个名字，如在A表中，关联B表的字段名是afieldb ，而在B表被A表关联的字段名是bfielda,则不能直接应用以上方式来进行联接查询,可以变通一下成这样：
return array(
'VarName'=>array('RelationType', 'ClassName', '', 'on' => ' t.afieldb = VarName.bfielda ',...additional options)
);
解释一下：

array的前三个元素是不能变了，他们的意义与他们的顺序相对应，第一个元素一定得是关联类型，第二个元素得是被关联表model的类名，第三个元素得
是关联字段，除些三个，其它元素则与位置关系不大了，不过，必须指定元素的key，如本例中的'on'就是键名，这里的第三个元素是空，就两个单引号，后
面'on'的数据键值指定他们的关联关系，t.afieldb，表示主表字段afieldb，VarName.bfielda表示被关联表字段
bfielda,顺便说明一下，在所有relations关系中，主表都有个别名，并且这个别名就是t,而被关联表也有一个别名，这个别名就是定义的
VarName，如：
'shops'=>array(self::HAS_ONE, 'SubjectShops', '' , 'on'=>' t.sid = shops.id '),
其中，被关联表别名就是shops。

除了这个'on'元素外，在关系声明时可以指定附加的选项,概括如下：
select: 关联的 AR 类中要选择(select)的列的列表。 默认为 '*'，即选择所有列。此选项中的列名应该是已经消除歧义的。

condition: 即 WHERE 条件。默认为空。此选项中的列名应该是已经消除歧义的。

params: 要绑定到所生成的 SQL 语句的参数。应该以 名-值 对数组的形式赋值。此选项从 1.0.3 版起有效。

on: 即 ON 语句。此处指定的条件将会通过 AND 操作符附加到 join 条件中。此选项中的列名应该是已经消除歧义的。 此选项不会应用到 MANY_MANY 关系中。此选项从 1.0.2 版起有效。

order: 即 ORDER BY 语句。默认为空。 此选项中的列名应该是已经消除歧义的。

with:
a list of child related objects that should be loaded together with
this object. Be aware that using this option inappropriately may form an
infinite relation loop.

joinType: type of join for this relationship. It defaults to LEFT OUTER JOIN.

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 the same as the relation name.

together: whether
the table associated with this relationship should be forced to join
together with the primary table and other tables. This option is only
meaningful for HAS_MANY and MANY_MANY relations. If this option is set
false, the table associated with the HAS_MANY or MANY_MANY relation will
be joined with the primary table in a separate SQL query, which may
improve the overall query performance since less duplicated data is
returned. If this option is set true, the associated table will always
be joined with the primary table in a single SQL query, even if the
primary table is paginated. If this option is not set, the associated
table will be joined with the primary table in a single SQL query only
when the primary table is not paginated. For more details, see the
section "Relational Query Performance". This option has been available
since version 1.0.3.

join: the extra JOIN clause. It defaults to empty. This option has been available since version 1.1.3.

group: the GROUP BY clause. It defaults to empty. Column names referenced in this option should be disambiguated.

having:
the HAVING clause. It defaults to empty. Column names referenced in
this option should be disambiguated. Note: option has been available
since version 1.0.1.

index: the name of the column whose values
should be used as keys of the array that stores related objects. Without
setting this option, an related object array would use zero-based
integer index. This option can only be set for HAS_MANY and MANY_MANY
relations. This option has been available since version 1.0.7.

In addition, the following options are available for certain relationships during lazy loading:

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.