In this document

"Mediates between the domain and data mapping layers using a collection-like interface for accessing domain objects" (Martin Fowler).

Repositories, in practice, are used to perform database operations for domain objects (Entity and Value types). Generally, a seperated repository is used for each Entity (or Aggregate Root).

Default Repositories

In ASP.NET Boilerplate, a repository classes implement IRepository<TEntity, TPrimaryKey> interface. ABP can automatically creates default repositories for each entity type. You can directly inject IRepository<TEntity> (or IRepository<TEntity, TPrimaryKey>). An example application service uses a repository to insert an entity to database:

public class PersonAppService : IPersonAppService
{
    private readonly IRepository<Person> _personRepository;

    public PersonAppService(IRepository<Person> personRepository)
    {
        _personRepository = personRepository;
    }

    public void CreatePerson(CreatePersonInput input)
    {        
        person = new Person { Name = input.Name, EmailAddress = input.EmailAddress };
        _personRepository.Insert(person);
    }
}

PersonAppService contructor-injects IRepository<Person> and uses the Insert method.

Custom Repositories

You only create a repository class for an entity when you need to create a custom repository method(s) for that entity.

Custom Repository Interface

A repository definition for a Person entity is shown below:

public interface IPersonRepository : IRepository<Person>
{

}

IPersonRepository extends IRepository<TEntity>. It's used to define entities which has a primary key type of int (Int32). If your entity's primary key is not int, you can extend IRepository<TEntity, TPrimaryKey> interface as shown below:

public interface IPersonRepository : IRepository<Person, long>
{

}

Custom Repository Implementation

ASP.NET Boilerplate is designed to be independent from a particular ORM (Object/Relational Mapping) framework or another technique to access to database. Repositories are implemented in NHibernate and EntityFramework as out-of-the-box. See documents to implement repositories in ASP.NET Boilerplate in these frameworks:

Base Repository Methods

Every repository has some common methods coming from IRepository<TEntity> interface. We will investigate most of them here.

Querying

Getting single entity
TEntity Get(TPrimaryKey id);
Task<TEntity> GetAsync(TPrimaryKey id);
TEntity Single(Expression<Func<TEntity, bool>> predicate);
Task<TEntity> SingleAsync(Expression<Func<TEntity, bool>> predicate);
TEntity FirstOrDefault(TPrimaryKey id);
Task<TEntity> FirstOrDefaultAsync(TPrimaryKey id);
TEntity FirstOrDefault(Expression<Func<TEntity, bool>> predicate);
Task<TEntity> FirstOrDefaultAsync(Expression<Func<TEntity, bool>> predicate);
TEntity Load(TPrimaryKey id);

Get method is used to get an Entity with given primary key (Id). It throws exception if there is no entity in database with given Id. Single method is similar to Get but takes an expression rather than an Id. So, you can write a lambda expression to get an Entity. Example usages:

var person = _personRepository.Get(42);
var person = _personRepository.Single(p => p.Name == "Halil İbrahim Kalkan");

Notice that Single method throws exception if there is no entity with given conditions or there are more than one entity.

FirstOrDefault is similar but returns null (instead of throwing exception) if there is no entity with given Id or expression. Returns first found entity if there are more than one entity for given conditions.

Load does not retrieves entity from database but creates a proxy object for lazy loading. If you only use Id property, Entity is not actually retrieved. It's retrieved from database only if you access to other properties of entity. This can be used instead of Get, for performance reasons. It's implemented in NHibernate. If ORM provider does not implements it, Load method works as identical as Get method.

Getting a list of entities
List<TEntity> GetAllList();
Task<List<TEntity>> GetAllListAsync();
List<TEntity> GetAllList(Expression<Func<TEntity, bool>> predicate);
Task<List<TEntity>> GetAllListAsync(Expression<Func<TEntity, bool>> predicate);
IQueryable<TEntity> GetAll();

GetAllList is used to retrieve all entities from database. Overload of it can be used to filter entities. Examples:

var allPeople = _personRepository.GetAllList();
var somePeople = _personRepository.GetAllList(person => person.IsActive && person.Age > 42);

GetAll returns IQueryable<T>. So, you can add Linq methods after it. Examples:

//Example 1
var query = from person in _personRepository.GetAll()
            where person.IsActive
            orderby person.Name
            select person;
var people = query.ToList();

//Example 2:
List<Person> personList2 = _personRepository.GetAll().Where(p => p.Name.Contains("H")).OrderBy(p => p.Name).Skip(40).Take(20).ToList();

With using GetAll, almost all queries can be written in Linq. Even it can be used in a join expression.

About IQueryable<T>

When you call GetAll() out of a repository method, there must be an open database connection. This is because of deferred execution of IQueryable<T>. It does not perform database query unless you call ToList() method or use the IQueryable<T> in a foreach loop (or somehow access to queried items). So, when you call ToList() method, database connection must be alive. For a web application, you don't care about that in most cases since MVC controller methods are unit of work by default and database connection is available for entire request. See UnitOfWork documentation to understand it better.

Custom return value

There is also an additional method to provide power of IQueryable that can be usable out of a unit of work.

T Query<T>(Func<IQueryable<TEntity>, T> queryMethod);

Query method accepts a lambda (or method) that recieves IQueryable<T> and returns any type of object. Example:

var people = _personRepository.Query(q => q.Where(p => p.Name.Contains("H")).OrderBy(p => p.Name).ToList());

Since given lamda (or method) is executed inside the repository method, it's executed when database connection is available. You can return a list of entities, a single entity, a projection or something else that executes the query.

Insert

IRepository interface defines methods to insert an entity to database:

TEntity Insert(TEntity entity);
Task<TEntity> InsertAsync(TEntity entity);
TPrimaryKey InsertAndGetId(TEntity entity);
Task<TPrimaryKey> InsertAndGetIdAsync(TEntity entity);
TEntity InsertOrUpdate(TEntity entity);
Task<TEntity> InsertOrUpdateAsync(TEntity entity);
TPrimaryKey InsertOrUpdateAndGetId(TEntity entity);
Task<TPrimaryKey> InsertOrUpdateAndGetIdAsync(TEntity entity);

Insert method simply inserts new entity to database and returns the same inserted entity. InsertAndGetId method returns Id of new inserted entity. This is useful if Id is auto increment and you need Id of the new inserted entity. InsertOrUpdate inserts or updated given entity by checking it's Id's value. Lastly, InsertOrUpdateAndGetId returns Id of the entity after inserting or updating.

Update

IRepository defines methods to update an existing entity in the database. It gets the entity to be updated and returns the same entity object.

TEntity Update(TEntity entity);
Task<TEntity> UpdateAsync(TEntity entity);

Most of times you don't need to explicitly call Update methods since unit of work system automatically saves all changes when unit of work completes. See unit of work documentation for more.

Delete

IRepository defines methods to delete an existing entity from the database

void Delete(TEntity entity);
Task DeleteAsync(TEntity entity);
void Delete(TPrimaryKey id);
Task DeleteAsync(TPrimaryKey id);
void Delete(Expression<Func<TEntity, bool>> predicate);
Task DeleteAsync(Expression<Func<TEntity, bool>> predicate);

First method accepts an existing entity, second one accepts Id of the entity to delete. The last one accepts a condition to delete all entities fit to given condition. Notice that all entities matches given predicate may be retrived from database and then deleted (based on repository implementation). So, use it carefully, it may cause performance problems if there are too many entities with given condition.

Others

IRepository also provides methods to get count of entities in a table.

int Count();
Task<int> CountAsync();
int Count(Expression<Func<TEntity, bool>> predicate);
Task<int> CountAsync(Expression<Func<TEntity, bool>> predicate);
long LongCount();
Task<long> LongCountAsync();
long LongCount(Expression<Func<TEntity, bool>> predicate);
Task<long> LongCountAsync(Expression<Func<TEntity, bool>> predicate);

About Async Methods

ASP.NET Boilerplate supports async programming model. So, repository methods has Async versions. Here, a sample application service method that uses async model:

public class PersonAppService : AbpWpfDemoAppServiceBase, IPersonAppService
{
    private readonly IRepository<Person> _personRepository;

    public PersonAppService(IRepository<Person> personRepository)
    {
        _personRepository = personRepository;
    }

    public async Task<GetPeopleOutput> GetAllPeople()
    {
        var people = await _personRepository.GetAllListAsync();

        return new GetPeopleOutput
        {
            People = Mapper.Map<List<PersonDto>>(people)
        };
    }
}

GetAllPeople method is async and uses GetAllListAsync with await keyword.

Async may not be supported by all ORM frameworks. It's supported by EntityFramework. If not supported, Async repository methods works synchronously. Also, for example, InsertAsync works same as Insert in EntityFramework since EF does not write new entities to database until unit of work completes (a.k.a. DbContext.SaveChanges).

Managing Database Connection

A database connection is not opened or closed in a repository method. Connection management is made automatically by ASP.NET Boilerplate.

A database connection is opened and a transaction begins while entering a repository method automatically. When the method ends and returns, all changes are saved, transaction is commited and database connection is closed automatically by ASP.NET Boilerplate. If your repository method throws any type of Exception, the transaction is automatically rolled back and database connection is closed. This is true for all public methods of classes those implement IRepository interface.

If a repository method calls to another repository method (even a method of different repository) they share same connection and transaction. Connection is managed (opened/closed) by the first method that enters a repository. For more information on database connection management, see UnitOfWork documentation.

Lifetime of a Repository

All repository instances are Transient. It means, they are instantiated per usage. See Dependency Injection documentation for more information.

Repository Best Practices

  • For an entity of T, use IRepository<T> wherever it's possible. Don't create custom repositories unless it's really needed. Pre-defined repository methods will be enough for many cases.
  • If you are creating a custom repository (by extending IRepository<TEntity>);
    • Repository classes should be stateless. That means, you should not define repository-level state objects and a repository method call should not effect another call.
    • Custom repository methods should not contain business logic or application logic. It should just perform data-related or orm-specific tasks.
    • While repositories can use dependency injection, define less or no dependency to other services.