可以为实体指定附加的限制条件，NHibernate 在执行 SELECTs 时会始终遵守这些条件。这些条件有两种:限制和过滤。让我们首先关注限制。
限制是一个静态 WHERE 子句，可以在类和集合级别(集合、列表、包、映射、 id 包、数组和原始数组)添加，以过滤其元素。一个典型的例子是软删除，其中记录实际上并没有从数据库中删除，而是用表示删除的列值来标记。下面是我们如何为类和集合将它定义为 XML:
<class name="Blog" lazy="true" table="`POST`" where="deleted = 0">
<!-- ... -->
<list cascade="all-delete-orphan" inverse="true" lazy="true" name="Posts" where="deleted = 0">
<!-- ... -->
</list>
</class>下面是代码映射中的等效代码:
mapper.Class<Blog>(ca =>
{
ca.Where("deleted = 0");
//...
ca.List(c => c.Posts, c =>
{
c.Where("deleted = 0");
//...
});
}在属性中:
[Class(Table = "blog", Lazy = true, Where = "deleted = 0")]
public class Blog
{
//...
[List(0, Inverse = true, Lazy = CollectionLazy.True, Generic = true, Where = "deleted = 0", Cascade = "all-delete-orphan")]
public virtual IList<Post> Posts { get; protected set; }
}deleted = 0”限制:
IEnumerable<Blog> nonDeletedBlogs = session.Query<Blog>();
Blog b = session.Get<Blog>(1);查看两个查询的结果 SQL:
SELECT
blog0_.blog_id AS blog1_11_,
blog0_.picture AS picture11_,
blog0_.user_id AS user3_11_,
blog0_.name AS name11_,
blog0_.creation AS creation11_,
(SELECT
COUNT(1)
FROM
post
WHERE
post.blog_id = blog0_.blog_id) AS formula2_
FROM
blog blog0_
WHERE
(
blog0_.deleted = 0
)
SELECT
blog0_.blog_id AS blog1_11_0_,
blog0_.picture AS picture11_0_,
blog0_.user_id AS user3_11_0_,
blog0_.name AS name11_0_,
blog0_.creation AS creation11_0_,
(SELECT
COUNT(1)
FROM
post
WHERE
post.blog_id = blog0_.blog_id) AS formula2_0_
FROM
blog blog0_
WHERE
blog0_.blog_id = 1
AND
xml:lang="PT">(
blog0_.deleted = 0
)过滤器类似于限制，但它是动态的。这意味着它可以启用或禁用，并且可以有参数。一个例子可能是您将术语翻译成多种语言的模型:
图 25:翻译类模型
我们所拥有的是:
- 一种语言
- 一个术语和一组翻译
- 给定语言的术语的翻译
这引入了一个新概念:映射复合键的类。在这种情况下，是语言术语包含了对术语和语言的引用。
通常，如果您只想加载当前语言的翻译，而不是所有语言的翻译，您会使用这样的模型。它的映射可能如下所示:
public class LanguageMapping : ClassMapping<Language>
{
public LanguageMapping()
{
this.Table("language");
this.Lazy(true);
this.Id(x => x.LanguageId, x =>
{
x.Column("language_id");
x.Generator(Generators.Assigned);
});
this.Property(x => x.Name, x =>
{
x.Column("name");
x.NotNullable(true);
x.Length(100);
});
}
} public class TermMapping : ClassMapping<Term>
{
public TermMapping()
{
this.Table("term");
this.Lazy(true);
this.Id(x => x.TermId, x =>
{
x.Column("term_id");
x.Generator(Generators.HighLow);
});
this.Property(x => x.Description, x =>
{
x.Column("description");
x.NotNullable(true);
x.Length(50);
});
this.Set(x => x.Translations, x =>
{
x.Key(y =>
{
y.Column("term_id");
y.NotNullable(true);
});
x.Filter("CurrentLanguage", z =>
{
z.Condition("language_id = :code");
});
x.Inverse(true);
x.Cascade(Cascade.All | Cascade.DeleteOrphans);
x.Lazy(CollectionLazy.Lazy);
}, x =>
{
x.OneToMany();
});
}
} public class TranslationMapping : ClassMapping<Translation>
{
public TranslationMapping()
{
this.Table("translation");
this.Lazy(true);
this.Filter("CurrentLanguage", x =>
{
x.Condition("language_id = :code");
});
this.ComponentAsId(x => x.TranslationId, x =>
{
x.ManyToOne(y => y.Language, y =>
{
y.Column("language_id");
});
x.ManyToOne(y => y.Term, y =>
{
y.Column("term_id");
});
});
this.Property(x => x.Text, x =>
{
x.Column("text");
x.Length(100);
x.NotNullable(true);
});
}
}最后但并非最不重要的是，在会话工厂构建之前，必须在配置实例上进行过滤器声明:
cfg.AddFilterDefinition(new FilterDefinition("CurrentLanguage", "language_id = :code",
new Dictionary<String, IType>() { { "code", NHibernateUtil.String } }, false));| 
| 提示:导入命名空间 NHibernate.Engine。 |
请注意:
- CurrentLanguage 过滤器基本上是对 language_id 列的限制，它使用了一个代码参数。
- 语言类的主键语言标识，是一个字符串，使用指定的模式。我们将使用它来存储有意义的文化名称，如“ en-us ”或“ pt-pt ”，它们必须始终是唯一的。
- 术语类的翻译集合应用了当前语言过滤器。
- 翻译类有一个由语言术语类实现的复合键(组件，它引用了一个术语和一个语言。
- 翻译类也使用货币语言过滤器。
我们需要给过滤器的代码参数赋值，并在实际查询前启用它:
//set the filter value from the current thread’s culture name
session.EnableFilter("CurrentLanguage").SetParameter("code", Thread.CurrentThread.CurrentCulture.Name);
var term = session.Query<Term>().First();
>//the translations will be filtered
var translations = term.Translations.ToList();过滤器将作为参数包含在 SQL 中:
SELECT
TOP (1) term0_.term_id AS term1_8_,
term0_.description AS descript2_8_
FROM
term term0_
SELECT
translatio0_.term_id AS term2_1_,
translatio0_.language_id AS language1_1_,
translatio0_.language_id AS language1_9_0_,
translatio0_.term_id AS term2_9_0_,
translatio0_.text AS text9_0_
FROM
[translation] translatio0_
WHERE
translatio0_.language_id = 'en-us'
AND translatio0_.term_id = 1当您不再需要过滤器时，只需将其禁用:
//disable the filter
session.DisableFilter("CurrentLanguage");最终注释:
- 每个实体和集合可能有多个筛选器，并且可以同时启用多个筛选器。
- 过滤器可以有参数，也可以没有参数，这些参数可以有默认值。
- 可以启用或禁用过滤器。
- 筛选器的 SQL 只能引用应用它的类可直接访问的列。