I'm using DataNucleus 2 JDO implementation. I have a detached object that i must attach, BUT i don't want to attach ALL the fields (in this case, a collection)
public class Obj {
private String key;
private Collection<String> col;
}
Is there a reason why it is not possible to do this:
tx.begin();
obj.makeTransientAll(obj.getCol()); // Do not persist
pm.makePersistent(obj);
tx.commit();
or REFRESH from the database:
tx.begin();
obj.refreshAll(obj.getCol()); // Discard any changes
pm.makePersistent(obj);
tx.commit();
Thanks.
I guess you should apply your changes AFTER refreshing your object or making it transient.
Or on the other hand, do something I really wanted to do to understand more how JDO works but I didn't have much time to investigate. Why don't you after initially looking up your object, do not detach it, but keep it in attached state and modify outside the look up method, then make it persistence in another method ?
What confuses me here is that in the look up method, you'll not close your PersistenceManager and in the latter method that saves your object you'll be using another PersistenceManager which is ought to fire an exception because you are managing an object using a different PersistenceManager other than the one that looked it up on the first place.
I'm still learning about JDO and Datanucleus so please expect that my suggest may not work. And sharing your experience with my suggestion is very much appreciated.
Related
Apologies, in advance, if this seems like a duplicate question. This question was the closest I could find, but it doesn't really solve the issues I am facing.
I'm using Entity Framework 5 in an ASP.NET MVC4 application and attempting to implement the Unit of Work pattern.
My unit of work class implements IDisposable and contains a single instance of my DbContext-derived object context class, as well as a number of repositories, each of which derives from a generic base repository class that exposes all the usual repository functionality.
For each HTTP request, Ninject creates a single instance of the Unit of Work class and injects it into the controllers, automatically disposing it when the request is complete.
Since EF5 abstracts away the data storage and Ninject manages the lifetime of the object context, it seems like the perfect way for consuming code to access in-memory entity objects without the need to explcitly manage their persistence. In other words, for optimum separation of concerns, I envisage my controller action methods being able to use and modify repository data without the need to explicitly call SaveChanges afterwards.
My first (naiive) attempt to implement this idea employed a call to SaveChanges within every repository base-class method that modified data. Of course, I soon realized that this is neither performance optimized (especially when making multiple successive calls to the same method), nor does it accommodate situations where an action method directly modifies a property of an object retrieved from a repository.
So, I evolved my design to eliminate these premature calls to SaveChanges and replace them with a single call when the Unit of Work instance is disposed. This seemed like the cleanest implementation of the Unit of Work pattern in MVC, since a unit of work is naturally scoped to a request.
Unfortunately, after building this concept, I discovered its fatal flaw - the fact that objects added to or deleted from a DbContext are not reflected, even locally, until SaveChanges has been called.
So, what are your thoughts on the idea that consuming code should be able to use objects without explicitly persisting them? And, if this idea seems valid, what's the best way to achieve it with EF5?
Many thanks for your suggestions,
Tim
UPDATE: Based on #Wahid's response, I am adding below some test code that shows some of the situations in which it becomes essential for the consuming code to explicitly call SaveChanges:
var unitOfWork = _kernel.Get<IUnitOfWork>();
var terms = unitOfWork.Terms.Entities;
// Purge the table so as to start with a known state
foreach (var term in terms)
{
terms.Remove(term);
}
unitOfWork.SaveChanges();
Assert.AreEqual(0, terms.Count());
// Verify that additions are not even reflected locally until committed.
var created = new Term { Pattern = "Test" };
terms.Add(created);
Assert.AreEqual(0, terms.Count());
// Verify that additions are reflected locally once committed.
unitOfWork.SaveChanges();
Assert.AreEqual(1, terms.Count());
// Verify that property modifications to entities are reflected locally immediately
created.Pattern = "Test2";
var another = terms.Single(term => term.Id == created.Id);
Assert.AreEqual("Test2", another.Pattern);
Assert.True(ReferenceEquals(created, another));
// Verify that queries against property changes fail until committed
Assert.IsNull(terms.FirstOrDefault(term => term.Pattern == "Test2"));
// Verify that queries against property changes work once committed
unitOfWork.SaveChanges();
Assert.NotNull(terms.FirstOrDefault(term => term.Pattern == "Test2"));
// Verify that deletions are not even reflected locally until committed.
terms.Remove(created);
Assert.AreEqual(1, terms.Count());
// Verify that additions are reflected locally once committed.
unitOfWork.SaveChanges();
Assert.AreEqual(0, terms.Count());
First of all SaveChanges should NOT be ever in the repositories at all. Because that's leads you to lose the benefit of UnitOfWork.
Second you need to make a special method to save changes in the UnitOfWork.
And if you want to call this method automatically then you may fine some other solution like ActionFilter or maybe by making all your Controllers inherits from BaseController class and handle the SaveChanges in it.
Anyway the UnitOfWork should always have SaveChanges method.
We have a scenario in our code when only a few properties of an entity are allowed to be changed. To guarantee that, we have code similar to this:
public void SaveCustomer(Customer customer)
{
var originalCustomer = dbContext.GetCustomerById(customer.Id);
if (customer.Name != originalCustomer.Name)
{
throw new Exception("Customer name may not be changed.");
}
originalCustomer.Address = customer.Address;
originalCustomer.City = customer.City;
dbContext.SaveChanges();
}
The problem with this code is that the call to dbContext.GetCustomerById does not always gives me a new instance of the Customer class. If the customer already has been fetched from the database, Entity Framework will keep the instance in memory and return it on every subsequent call.
This leads us to the actual problem - customer and originalCustomer may refer to the same instance. In that case, customer.Name will be equal to originalCustomer.Name and we will not be able to detect if it differs from the database.
I guess the same problem exists with most other ORMs as well, because of the identitymap design pattern.
Any ideas how this can be solved? Can I somehow force EF to always give me a new instance of the customer class?
Or should we refactor the code instead? Does anyone know of any good design patterns for this scenario?
you can try by detaching the entity from the context, this will remove all the references to the context (as well as the identitymap behaviour).
So, before passing the Customer to your method you can detach it:
yourContext.Detach(customer);
All the examples I've found refer to a class called ObjectContext, which doesn't appear to exist in CTP5. I must stress at this point, CTP5 is my first exposure to the Entity Framework.
I have a disconnected POCO that I have attached to my DbContext. SaveChanges does not pick up the change though, how I tell my context to update that entity?
_context.Users.Attach(user);
// The user has been replaced.
_context.SaveChanges();
// The change is not saved.
What am I doing wrong?
Update 12/01/2011
Might be obvious to most, but as a first time user of EF, it didn't occur to me that attaching an object that was already attached would clear the previous state. This caused me a lot of pain. But I wanted to use the Repository pattern in a very generic way, a way which didn't care if the object was already attached or had been freshly created as the result of ASP.NET MVC binding. So I needed an UpdateUser method, and I've attached it below.
public User UpdateUser(User user) {
if (_context.Entry(user).State == EntityState.Detached) {
_context.Users.Attach(user);
_context.Entry(user).State = EntityState.Modified;
}
return user;
}
The method obviously assumes that the object exists in the data store in some fashion, it's called UpdateUser after all. If the object is already attached, you will benefit from the object's previous state, which in turn will allow for an optimised update to the DB. However, if the object was not attached, the method forces the whole thing to become dirty.
Seems obvious now, wasn't before. Hope it helps someone.
Rich
When you Attach an entity, it goes to Unchanged state (it has not been changed since it attached to the context). All you need to is to explicitly change the Entity State to Modified:
_context.Users.Attach(user);
_context.Entry(user).State = System.Data.Entity.EntityState.Modified;
_context.SaveChanges();
For the sake of completeness, you can access the ObjectContext by casting the DbContext to IObjectContextAdapter:
((IObjectContextAdapter)context).ObjectContext.ObjectStateManager.ChangeObjectState(user, EntityState.Modified);
Morteza's method is much cleaner though and gets my vote.
I believe u do not need to attach the entity before u call modified. simply setting to modified will do the job.
if (_context.Entry(user).State == EntityState.Detached)
{
_context.Entry(user).State = EntityState.Modified;
}
Ive just started using EF4 with the repository pattern. Im having to call the dispose method after every use of context or wrap code arround in the using block. Can I use the ObjectContext without doing this in every method I write or is there a better way of handling this in the repository.
Also I dont want to pass the ObjectContext to the repository from the UI as well.
To do this as resource effectively as possible without dependency injection, I would suggest that you implement a private, lazy-loading property for the object context.
private ObjectContext _context;
private ObjectContext Context
{ get
{
return _context ?? (_context = new ObjectContext());
}
}
Next, make your repository implement IDisposable and take care of the object context in your dispose method:
public Repository : IDisposable
{
...
public void Dispose()
{
_context.Dispose();
}
}
Then, you just use the property in all your methods, and wrap usage of your repository in using statements.
To decrease traffic to the database, you could also factor out the saving to a separate method on the repository, which just forwards the call to the object context. That way you get more control of when data is saved in the UI layer, even though you don't control how. That means you could do
using (var repo = new Repository())
{
repo.AddSomeStuff("this", "is", true);
repo.ChangeSomethingElse("yes, please");
repo.Save();
}
and there would only be one call from EF to the database. On the other hand, if you do
using (var repo = new Repository())
{
repo.AddSomeStuff("this", "is", true);
repo.ChangeSomethingElse("yes, please");
}
nothing happens, which might be confusing.
The general pattern for using an object context is:
public BusinessObject GetSomething(){
using (MyObjectContext context = new MyObjectContext()){
//..do fun stuff
}
}
Which hopefully is the pattern you are using. Calling dispose seems a little overkill when you can just use a "using" statement.
Another option is if you are going to be doing multiple DB queries in a flow. I have seen a pattern where you can reuse the same context within the thread. People basically implement a Thread based singleton pattern and pass around the context. Advantages to this is not having to rebuild the context, as well as some in memory caching. Downside is you could run into concurrency issues. Someone updating something that you have cached internally in EF.
I am guessing the second case doesn't really apply because it sounds like you are writting a small app. (that statement was based on your comments about passing a context from UI...a statement which will scare any good code architect).
If you are interested in a thread based singleton. First learn about the Singleton pattern, and then check out this blog about "DataContext" threads. You will have to change the "DataContext" type to the ObjectContext class but it would work.
EDIT
I will say I overlooked an obvious solution and that would be the below ;). Just using a property based Object Context and playing your repository in a using statement. It would be the same as the using example above, but you would implement IDisoposable.
I'm a total newbie at Entity Framework and ASP.Net MVC, having learned mostly from tutorials, without having a deep understanding of either. (I do have experience on .Net 2.0, ADO.Net and WebForms)
My current doubt comes from the way I'm instancing my Entities objects.
Basically I'm doing this in my controllers:
public class PostsController : Controller {
private NorthWindEntities db = new NorthWindEntities();
public ActionResult Index() {
// Use the db object here, never explicitly Close/Dispose it
}
}
I'm doing it like this because I found it in some MSDN blog that seemed authoritative enough to me that I assumed this was a correct way.
However, I feel pretty un-easy about this. Although it saves me a lot of code, I'm used to doing:
using (NorthWindEntities db = new NorthWindEntities() {
}
In every single method that needs a connection, and if that method calls others that'll need it, it'll pass db as a parameter to them. This is how I did everything with my connection objects before Linq-to-SQL existed.
The other thing that makes me uneasy is that NorthWindEntities implements IDisposable, which by convention means I should be calling it's Dispose() method, and I'm not.
What do you think about this?
Is it correct to instance the Entities object as I'm doing? Should it take care of its connections by opening and closing them for each query?
Or should I be disposing it explicitly with a using() clause?
Thanks!
Controller itself implements IDisposable. So you can override Dispose and dispose of anything (like an object context) that you initialize when the controller is instantiated.
The controller only lives as long as a single request. So having a using inside an action and having one object context for the whole controller is exactly the same number of contexts: 1.
The big difference between these two methods is that the action will have completed before the view has rendered. So if you create your ObjectContext in a using statement inside the action, the ObjectContext will have been disposed before the view has rendered. So you better have read anything from the context that you need before the action completes. If the model you pass to the view is some lazy list like an IQueryable, you will have disposed the context before the view is rendered, causing an exception when the view tries to enumerate the IQueryable.
By contrast, if you initialize the ObjectContext when the Controller is initialized (or write lazy initialization code causing it to be initialized when the action is run) and dispose of the ObjectContext in the Controller.Dispose, then the context will still be around when the view is rendered. In this case, it is safe to pass an IQueryable to the view. The Controller will be disposed shortly after the view is rendered.
Finally, I'd be remiss if I didn't point out that it's probably a bad idea to have your Controller be aware of the Entity Framework at all. Look into using a separate assembly for your model and the repository pattern to have the controller talk to the model. A Google search will turn up quite a bit on this.
You are making a good point here. How long should the ObjectContext live? All patterns and practises books (like Dino Esposito's Microsoft-NET-Architecting-Applications) tell you that a DataContext must not live long, nor should it be cached.
I was just wondering why not having, in your case, a ControllerBase class (I'm not aware of the MVC implementation, so bear with me) where the ObjectContext gets initiated once for all controller. Especially think about the Identity Map Pattern, that's already implemented by Entity Framework. Even though you need to call another controller as your PostsController, it would still work with the same Context and improve performance as well.