I have a repository like that:
public class Repository<T> : IRepository<T> where T : class
{
private readonly IRepositoryContext _repositoryContext;
public Repository(IRepositoryContext repositoryContext)
{
_repositoryContext = repositoryContext;
_objectSet = repositoryContext.GetObjectSet<T>();
}
public virtual void Update(T entity)
{
ObjectSet.AddObject(entity);
_repositoryContext.ObjectContext.ObjectStateManager.ChangeObjectState(entity, EntityState.Modified);
_repositoryContext.SaveChanges();
}
}
Now that actually works for all scalar properties of the entity, but all the other entities that associated with properties of entity typeOf(T), don't care that entity state is modified, and EF simply adds new data.
So, if you do for example Repository<Student>.Update(), and you only changed the name, it will find the right Student and change his name, but it also will change the Campus, although you already have a Campus associated with that student, it will be created again with a different CampusId.
Show me please the correct way to do updates in this situation.
What I did when I wanted to follow generic approach was translated to your code something like:
public class Repository<T> : IRepository<T> where T : class
{
...
public virtual void Update(T entity)
{
if (context.ObjectStateManager.GetObjectStateEntry(entity).State == EntityState.Detached)
{
throw new InvalidOperationException(...);
}
_repositoryContext.SaveChanges();
}
}
All my code then worked like:
var attachedEntity = repository.Find(someId);
// Merge all changes into attached entity here
repository.Update(attachedEntity);
=> Doing this in generic way moves a lot of logic into your upper layer. There is no better way how to save big detached object graphs (especially when many-to-many relations are involved and deleting of relations is involved).
Related
I would like to implement nlog to each action to add an element.
So when I do myContext.Society.Add(), I would like to log something.
I create a class DbSetExtension and modify the context StockContext to use DbSetExtension<T> instead DbSet.
public class DbSetExtension<T> : DbSet<T> where T : class
{
public override T Add(T entity)
{
LoggerInit.Current().Trace("Add Done");
return base.Add(entity);
}
}
When i launch the programm, I notice when I access to myContext.Society.Add.
Society is null. So I think I miss something with my class DbSetExtension but I don't find.
public class StockContext : DbContext
{
public StockContext()
: base("StockContext")
{
}
public DbSet<HistoricalDatas> HistoricalDatas { get; set; }
public DbSet<Society> Society { get; set; }
protected override void OnModelCreating(DbModelBuilder modelBuilder)
{
modelBuilder.Conventions.Remove<PluralizingTableNameConvention>();
}
}
Do you have any idea,
Regards,
Alex
[UPDATE]
Code allows to add.
If I replace DbSetExtension by DbSet, the same code works.
So my assumption is I miss something when I inherit from DbSet.
public bool SetSymbols()
{
CsvTools csvThreat = new CsvTools();
List<Eoddata> currentEnum =
csvThreat.ExtractData<Eoddata>(ConfigurationManager.GetString("FilePathQuotes", ""));
currentEnum.ForEach(
c =>
{
//LoggerInit.Current().Trace("Add Done");
Sc.Society.Add(
new Society()
{
RealName = c.Description,
Symbol = String.Format("{0}.PA", c.Symbol),
IsFind = !String.IsNullOrEmpty(c.Description)
});
});
if (Sc.SaveChanges() > 0)
return true;
return false;
}
In my opinion you took totally wrong direction. DbContext is made to work with DbSet and not DbSetExtension class. It is able to instantiate objects of type DbSet and not your own type. This is basically why you get this exception. Reparing it would require probably hacking EF internals and I fear that this problem will be just a beginning for you. Instead I would recommend you to use general way of logging with EF with use of interceptor classes. Here this is explained in details at the end of article Logging and Intercepting Database Operations. Generally this approach would be much more advantageous for you. Why? Because DbContext is just man-in-the-middle in communication with db. In logs you generally cares about what happens to db and its data. Calling Add method on DbSet may not have any effect at all if SaveChanges won't be called lated on. On contrary query interceptors lets you log strictly only interaction with db. Basing on query sent to db you may distinguish what is going on.
But if you instist on your approach I would recommend you using extension methods instead of deriving from DbSet:
public static class DbSetExtensions
{
public static T LoggingAdd<T>(this DbSet<T> dbSet, T entity)
{
LoggerInit.Current().Trace("Add Done");
return dbSet.Add(entity);
}
}
and call it like this:
context.Stock.LoggingAdd(entity);
I created a generic repository class that all my other repository classes are inheriting from. This is great, because it means almost all the plumbing is done one time for all repositories. I put a full explanation of what I'm talking about here, but here is the code for my GenericRepository (some code is removed for brevity):
public abstract class GenericRepository<T> : IGenericRepository<T> where T : class, new()
{
private IMyDbContext _myDbContext;
public GenericRepository(IMyDbContext myDbContext)
{
_myDbContext = myDbContext;
}
protected IMyDbContext Context
{
get
{
return _myDbContext;
}
}
public IQueryable<T> AsQueryable()
{
IQueryable<T> query = Context.Set<T>();
return query;
}
public virtual void Create(T entity)
{
Context.Set<T>().Add(entity);
}
public virtual void Update(T entity)
{
Context.Entry(entity).State = System.Data.EntityState.Modified;
}
}
As you see, I have a Create method and an Update method. It would be very convenient to have a "CreateOrUpdate" method, so I don't have to manually check for existing objects each time I have to save something to the database.
Each of my objects in Entity Framework have an "Id", but the challenge here is that the GenericRepository works with "T".
Now, with that rather long introduction, to my specific question.
How do I create a generic CreateOrUpdate method for my GenericRepository?
UPDATE
After Marcins response, I implemented the following generic methods in my GenericRepository. It will take some time before I can test that it works as expected, but it looks very promising.
public virtual bool Exists(Guid id)
{
return Context.Set<T>().Any(t => t.Id == id);
}
public virtual void CreateOrUpdate(T entity)
{
if (Exists(entity.Id))
{
var oldEntity = GetSingle(entity.Id);
Context.Entry(oldEntity).CurrentValues.SetValues(entity);
Update(oldEntity);
}
else
{
Create(entity);
}
}
The code above has no less than 3 roundtrips to the database when updating. I'm sure it can be optimized, but it wasn't really the exercise for this question.
This question handles that topic better:
An object with the same key already exists in the ObjectStateManager. The ObjectStateManager cannot track multiple objects with the same key
Create a interface with Id property, implement it on every of your entities and add another generic constraint to your class:
public interface IEntity
{
int Id { get; set;}
}
And
public abstract class GenericRepository<T> : IGenericRepository<T> where T : class, IEntity, new()
With that, you'll be able to use Id property within your generic repository class.
Of course - Id don't have to be an int, it can be Guid as well.
Hoping someone could clear things up. In the following ViewModel, does using Entity Framework as my model eliminate the need to use [Model] and [[ViewModelToModel(...)] attributes? The code runs the same with or without them, because the binding in the view ignores them and binds to the ObservableCollection.
Comments?
public class MainWindowViewModel : ViewModelBase
{
Models.OneHour_DataEntities ctx;
public MainWindowViewModel()
: base()
{
Save = new Command(OnSaveExecute, OnSaveCanExecute);
ctx = new Models.OneHour_DataEntities();
Customers = new ObservableCollection<Models.Customer>(ctx.Customers);
}
public ObservableCollection<Models.Customer> Customers
{
get { return GetValue<ObservableCollection<Models.Customer>>(CustomersProperty); }
set { SetValue(CustomersProperty, value); }
}
public static readonly PropertyData CustomersProperty = RegisterProperty("Customers", typeof(ObservableCollection<Models.Customer>), null);
public Command Save { get; private set; }
private bool OnSaveCanExecute()
{
return true;
}
private void OnSaveExecute()
{
ctx.SaveChanges();
}
}
Catel uses different interfaces to take advantage of the models. For example, it uses the following interfaces:
IEditableObject => undoing changes to model when user cancels
INotifyPropertyChanged => update view model when model updates
If your entity model implements these interfaces, you can define a property as a model.
In your example however, you use an ObservableCollection (thus a list of models) as a model. That is not supported (or, again, the collection must support IEditableObject and INotifyPropertyChanged).
I am using EF 5.0 and the model first approach. I have build a GenericRepository that has the basic get, insert, delete etc statements. Like:
public virtual void Insert(TEntity entity)
{
dbSet.Add(entity);
}
My EF entities all have the attributes Modified and ModifiedBy. Now I want to change this values everytime I save an entity.
Is it possible to modify this two attributes (set the value) without writing an specific implementation all the time?
Thank you
I see two options for you to do this, but they both entail either introducing a base type or an interface for all of your entities to cover them in a generic function. I would prefer an interface, although each entity would have to implement it again and again.
Let's say you create
interface IAuditable
{
DateTime Modified { get; set; }
string ModifiedBy {get; set; } // User id?
}
Now you can do:
public virtual void Insert(TEntity entity)
where TEntity : IAuditable
{
entity.Modified = DateTime.Now;
entity.ModifiedBy = ???? // Whatever you get the name from
...
}
(Same for edit)
You can also subscribe to the context's SavingChanges event:
// In the constructor:
context.SavingChanges += this.context_SavingChanges;
private void context_SavingChanges(object sender, EventArgs e)
{
foreach (var auditable in context.ObjectStateManager
.GetObjectStateEntries(EntityState.Added | EntityState.Modified)
.Select(entry => entry.Entity)
.OfType<IAuditable>)
{
auditable.Modified = DateTime.Now;
auditable.ModifiedBy = ????;
}
}
If you work with DbContext you can get to the event by
((IObjectContextAdapter)this).ObjectContext.SavingChanges
I'd like to add that more reliable time tracking can (and maybe should) be achieved by database triggers. Now you depend on a client's clock.
You can do this using the following code in your all methods of repository where you want to.
public virtual void Edit(TEntity entity)
{
entity.Modified=DateTime.Now;
entity.ModifiedBy=User.Identity.Name;
//Other saving to repository code
}
i have started learning entity framework CTP5 by writing a windows application.
i have two models (Unit and Good) as following:
public class Unit : BaseEntity
{
public Unit()
{
Goods = new List<Good>();
}
public string name { get; set; }
public virtual ICollection<Good> Goods { get; set; }
}
public class Good : BaseEntity
{
public Int64 code { get; set; }
public string name { get; set; }
public virtual Unit Unit { get; set; }
}
i'm using a repository inteface named IRepository as below :
public interface IRepository
{
BaseEntity GetFirst();
BaseEntity GetNext(Int32 id);
BaseEntity GetPrevoius(Int32 id);
BaseEntity GetLast();
BaseEntity GetById(Int32 id);
void Update(int id, BaseEntity newEntity);
void Delete(int id);
void Insert(BaseEntity entity);
int GetMaxId();
IList GetAll();
}
every model has its own repository but maybe it is better to use a generic repository of BaseEntity type. A reference of GoodRepository is made in GoodForm and appropriate object of Good type is made by Activator object in common form methods like Insert/Update/Delete... as below :
private void InsertButton_Click(object sender, EventArgs e)
{
Unit unit = goodRepo.GetUnitById(Convert.ToInt32(UnitIdTextBox.Text));
if (unit == null)
{
unit = new Unit { Id = goodRepo.GetUnitMaxId(), Name = "Gram" };
}
var good = Activator.CreateInstance<Good>();
good.Id = string.IsNullOrEmpty(IdTextBox.Text) ? goodRepo.GetMaxId() : Convert.ToInt32(IdTextBox.Text);
IdTextBox.Text = good.Id.ToString();
good.Name = NameTextBox.Text;
good.Description = DescriptionTextBox.Text;
good.Unit = unit;
goodRepo.Insert(good);
}
and GoodRepository.Insert method is :
public void Insert(Model.BaseEntity entity)
{
using (PlanningContext context = new PlanningContext())
{
context.Goods.Add(entity as Good);
int recordsAffected = context.SaveChanges();
MessageBox.Show("Inserted " + recordsAffected + " entities to the database");
}
}
My problem is SaveChanges() generate an error "Violation of PRIMARY KEY constraint" and says it can not inset duplicate key in object 'dbo.Units.
but if i move my context to the form which i build Good object and insert it there everything is fine.
Can anybody guid me how to solve this issue?
thank in advance
The source of your problem is here:
using (PlanningContext context = new PlanningContext())
{
context.Goods.Add(entity as Good);
//...
}
You are adding the Good entity to a newly created and therefore initially empty context. Now, if you add an entity to the context EF will add the whole object graph of related entities to the context as well, unless the related entities are already attached to the context. That means that good.Unit will be put into the context in Added state as well. Since you don't seem to have an autogenerated identity key for the Unit class, EF tries to insert the good.Unit into the DB with the same key which is already in the database. This causes the exception.
Now, you could ad-hoc fix this problem by attaching the Unit to the context before you add a new Good:
using (PlanningContext context = new PlanningContext())
{
context.Units.Attach((entity as Good).Unit);
context.Goods.Add(entity as Good);
//...
}
But I would better rethink the design of your repository. It's not a good idea to create a new context in every repository method. The context plays the role of a unit of work and a unit of work is usually more a container for many database operations which belong closely together and should be committed in a single database transaction.
So, operations like your InsertButton_Click method should rather have a structure like this:
using (var context = CreateSomehowTheContext())
{
var goodRepo = CreateSomehowTheRepo(context); // Inject this context
var perhapsAnotherRepo = CreateTheOtherRepo(context); // Inject same context
Unit unit = goodRepo.GetUnitById(Convert.ToInt32(UnitIdTextBox.Text));
// unit is now attached to context
// ...
good.Unit = unit;
goodRepo.Insert(good); // should use the injected context and only do
// context.Goods.Add(good);
// It doesn't add unit to the context since
// it's already attached
// ...
context.SaveChanges();
}
Here you are working only with one single context and the repositories will get this context injected (in the constructor for instance). They never create their own context internally.
I suspect it's because GetUnitMaxId is returning the same value more than once. Is Id an auto-incrementing identity column? If so, you shouldn't try to make any assumptions about what that value might be in code.
Even if it's not an auto-incrementing identity column, you can only be sure of it's value when all others have been committed to the DB.
As a general design pattern, try to avoid the need to refer to Ids in code before they've been stored. EF can help with this by exploiting navigation properties (inter-entity object references).