ActiveRecord 的 CRUD 操作基本上都是由 ActiveRecordBase 的静态方法完成。其实例(含子类型)方法，以及 ActiveRecordMediator 只是对这些静态方法的包装调用而已。当然，ActiveRecordBase 方法又是对 NHibernate Session 方法的包装。
我们用下面这个实体来简单演示一下常用方法的使用。
[ActiveRecord("Users")]
public class User : ActiveRecordBase<User>
{
private int id;
[PrimaryKey(PrimaryKeyType.Identity, Access=PropertyAccess.FieldCamelcase)]
public int Id
{
get { return id; }
}
private string name;
[Property(Unique=true, NotNull=true)]
public string Name
{
get { return name; }
set { name = value; }
}
}
为了观察 NHibernate 生成的 SQL 语句，我们在配置文件中添加 "show_sql" 配置。
<?xml version="1.0" encoding="utf-8" ?>
<activerecord>
<config>
<add key="hibernate.connection.driver_class" value="NHibernate.Driver.SqlClientDriver" />
<add key="hibernate.dialect" value="NHibernate.Dialect.MsSql2000Dialect" />
<add key="hibernate.connection.provider" value="NHibernate.Connection.DriverConnectionProvider" />
<add key="hibernate.connection.connection_string" value="server=(local);uid=sa;pwd=123456;database=test" />
<add key="hibernate.show_sql" value="true" />
</config>
</activerecord>
1. Create
Create 通过调用 NHibernate.ISession.Save() 方法完成实体类型的存储操作。方法调用完成后，我们就能立即获取 PrimaryKey ID。
User user = new User();
user.Name = "User";
user.Create();
Console.WriteLine(user.Id);
SQL
NHibernate: INSERT INTO Users (Name) VALUES (@p0); select SCOPE_IDENTITY()
@p0 = 'User'
2. Update
Update 通过调用 NHibernate.ISession.Update() 方法完成实体类型的更新操作。
User user2 = User.Find(1);
user2.Name = "User2";
user2.Update();
SQL
NHibernate: SELECT user0_.Id as Id0_, user0_.Name as Name0_ FROM Users user0_ WHERE user0_.Id=@p0
@p0 = '1'
NHibernate: UPDATE Users SET Name = @p0 WHERE Id = @p1
@p0 = 'User2'
@p1 = '1'
多数时候，我们先从数据库获取实例，然后再更新。但也可以直接创建实体对象，进行赋值更新。
// 先将上面的实体改成 Id {get; set;}
User user2 = new User();
user2.Id = 1;
user2.Name = "abcd";
user2.Update();
SQL
NHibernate: UPDATE Users SET Name = @p0 WHERE Id = @p1
@p0 = 'abcd'
@p1 = '1'
对比上下两次生成的 SQL 语句，会发现后一种方法更 "快" 一些。但更新的前提是必须知道所有实体属性的 "值"，通常这是做不到的，所以用处不大。
3. Save
Save 方法比较有趣，它调用的是 NHibernate.ISession.SaveOrUpdate()。也就是说，如果是 "新" 实体，它会 Create，否则 Update。
User user = new User();
user.Name = "User";
user.Save();
Console.WriteLine(user.Id);
User user2 = User.Find(user.Id);
user2.Name = "User2";
user2.Save();
Console.WriteLine(User.Find(user.Id).Name);
SQL
NHibernate: INSERT INTO Users (Name) VALUES (@p0); select SCOPE_IDENTITY()
@p0 = 'User'
NHibernate: SELECT user0_.Id as Id0_, user0_.Name as Name0_ FROM Users user0_ WHERE user0_.Id=@p0
@p0 = '1'
NHibernate: UPDATE Users SET Name = @p0 WHERE Id = @p1
@p0 = 'User2'
@p1 = '1'
NHibernate: SELECT user0_.Id as Id0_, user0_.Name as Name0_ FROM Users user0_ WHERE user0_.Id=@p0
@p0 = '1'
4. Delete
删除实体比较简单。
User user = new User();
user.Name = "User";
user.Save();
user.Delete();
SQL
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '1'
哎~~~ DeleteAll 还是老样子，要是想删除 1000 万个实体数据，我强烈建议你先去泡杯茶。
for (int i = 0; i < 10; i++)
{
User user = new User();
user.Name = "User" + i;
user.Save();
}
User.DeleteAll();
SQL
NHibernate: select user0_.Id as Id, user0_.Name as Name from Users user0_
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '1'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '2'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '3'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '4'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '5'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '6'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '7'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '8'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '9'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '10'
DeleteAll 还支持集合删除。
List<User> users = new List<User>();
for (int i = 0; i < 10; i++)
{
User user = new User();
user.Name = "User" + i;
user.Save();
users.Add(user);
}
User.DeleteAll(new int[] { users[1].Id, users[3].Id });
SQL
NHibernate: SELECT user0_.Id as Id0_, user0_.Name as Name0_ FROM Users user0_ WHERE user0_.Id=@p0
@p0 = '2'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '2'
NHibernate: SELECT user0_.Id as Id0_, user0_.Name as Name0_ FROM Users user0_ WHERE user0_.Id=@p0
@p0 = '4'
NHibernate: DELETE FROM Users WHERE Id = @p0
@p0 = '4'
5. Refresh
Refresh 的作用是从数据库刷新实体信息。
User user = new User();
user.Name = "User";
user.Save();
User user2 = User.Find(user.Id);
user2.Name = "xxxx";
user2.Save();
Console.WriteLine(user.Name);
user.Refresh();
Console.WriteLine(user.Name);
SQL
NHibernate: SELECT user0_.Id as Id0_, user0_.Name as Name0_ FROM Users user0_ WHERE user0_.Id=@p0
@p0 = '1'
6. Find
AR 提供了多种 Find 方法用来查找实体。
(1) PrimaryKey
for (int i = 0; i < 10; i++)
{
User user = new User();
user.Name = "User" + i;
user.Save();
}
User u = User.Find(3);
Console.WriteLine(u.Name);
我们还可以用 TryFind 代替 Find。这个方法在没找到实体时，不会触发异常。
User u = User.TryFind(500);
Console.WriteLine(u == null);
(2) Property
for (int i = 0; i < 10; i++)
{
User user = new User();
user.Name = "User" + i;
user.Save();
}
User[] us = User.FindAllByProperty("Name", "User3");
foreach (User u in us) Console.WriteLine(u.Name);
(3) NHibernate.Expression.ICriterion
NHibernate Criterion 是 HQL 的 "另外一种表现方式"，可用来组合多种条件。常用的有 LtExpression(<)、GeExpression(>=)、EqExpression(=)、GtExpression(>)、LeExpression(<=)、LikeExpression、AndExpression、BetweenExpression、NotExpression、NotNullExpression、NullExpression 等。还可以使用 NHibernate.Express.Order 进行排序。
for (int i = 0; i < 10; i++)
{
User user = new User();
user.Name = "User" + i;
user.Save();
}
//User[] users = User.FindAll(new Order[]{ new Order("Id", false) }, new InExpression("Id", new object[]{ 1, 3, 5 }));
//User[] users = User.FindAll(new EqExpression("Name", "User3"));
//User[] users = User.FindAll(new OrExpression(new EqExpression("Name", "User1"), new EqExpression("Name", "User3")));
//User u = User.FindFirst(new BetweenExpression("Id", 3, 6));
User u = User.FindOne(new EqExpression("Id", 3));
注意 FindOne 和 FindFirst 的区别，如果表达式返回的记录超过 1，FindOne 会触发异常。
7. Exists
Exists 可用来判断实体类型是否已经创建了对应的数据表，还可以通过主键判断实体是否存在，或者通过 hql 语句做出复杂的判断。
// 判断数据表是否存在
Console.WriteLine(User.Exists());
// 通过主键判断
Console.WriteLine(User.Exists<int>(3));
// HQL
Console.WriteLine(User.Exists("Name=?", "User3"));
Console.WriteLine(User.Exists("Name=? or Id=?", "User400", 4));
8. Query
Query 和 Find 的最大区别是它对 HQL 的支持。AR 提供了三个 Query 对象：SimpleQuery、ScalarQuery、CountQuery。
SimpleQuery: 主要用于获取一组信息，如实体数组、单个属性数组。
for (int i = 0; i < 10; i++)
{
User user = new User();
user.Name = "User" + i;
user.Save();
}
//SimpleQuery query = new SimpleQuery(typeof(User), "from User user where user.Name = ?", "User2");
//User[] users = (User[])User.ExecuteQuery(query);
//SimpleQuery query = new SimpleQuery(typeof(User), typeof(String), "select user.Name from User user where user.Name = ?", "User2");
//string[] names = (string[])User.ExecuteQuery(query);
//SimpleQuery<User> query = new SimpleQuery<User>("from User user where user.Name = ?", "User2");
//User[] users = query.Execute();
SimpleQuery<string> query = new SimpleQuery<string>(typeof(User), "select user.Name from User user where user.Name = ?", "User2");
string[] names = query.Execute();
ScalarQuery: 主要用于获取单个实体对象，或单个实体的多个属性。当返回实体记录数大于 1 时会触发异常。
for (int i = 0; i < 10; i++)
{
User user = new User();
user.Name = "User" + i;
user.Save();
}
//ScalarQuery query = new ScalarQuery(typeof(User), "from User user where user.Name = ?", "User2");
//User u = (User)User.ExecuteQuery(query);
//ScalarQuery query = new ScalarQuery(typeof(User), "select user.Id, user.Name from User user where user.Name = ?", "User2");
//object[] properties = (object[])User.ExecuteQuery(query);
//ScalarQuery<User> query = new ScalarQuery<User>(typeof(User), "from User user where user.Name = ?", "User2");
//User u = query.Execute();
ScalarQuery<object> query = new ScalarQuery<object>(typeof(User), "select user.Id, user.Name from User user where user.Name = ?", "User2");
object properties = query.Execute();
CountQuery: 用于统计数据库中实体记录数量。
//CountQuery query = new CountQuery(typeof(User));
//Console.WriteLine(User.ExecuteQuery(query));
CountQuery query = new CountQuery(typeof(User), "Name like ?", "User1%");
Console.WriteLine(User.ExecuteQuery(query));
我们还可以使用重载方法中的 QueryLanguage 参数，直接执行 SQL 语句，不过要谨慎使用，因为这样一来，就有可能失去自由切换数据库的能力。
9. Paging
利用 CountQuery 和 ActiveRecordBase.SlicedFindAll() 我们可以实现分页查询功能。
for (int i = 1; i <= 100; i++)
{
User user = new User();
user.Name = "User" + i.ToString().PadLeft(3, '0');
user.Save();
}
int pageIndex = 5;
int pageSize = 10;
CountQuery countQuery = new CountQuery(typeof(User), "Id < ?", 45);
int count = (int)User.ExecuteQuery(countQuery);
if (count > 0)
{
int pageCount = count / pageSize + (count % pageSize > 0 ? 1 : 0);
pageIndex = Math.Min(pageIndex, pageCount);
int first = Math.Max((pageIndex - 1) * pageSize, 0);
User[] users = User.SlicedFindAll(first, pageSize, new LtExpression("Id", 45));
foreach (User u in users) Console.WriteLine(u.Name);
}
SQL
NHibernate: select COUNT(*) as x0_0_ from Users user0_ where (Id<@p0)
@p0 = '45'
NHibernate: SELECT top 50 this.Id as Id0_, this.Name as Name0_ FROM Users this WHERE this.Id<@p0
@p0 = '45'
看它生成的 SQL 语句，可见其效率不是很好。只是不知道使用 SQLite / MySQL 时，是否会使用 LIMIT 语句进行分页处理。有空研究一下……
