MVC 3 + EF 4.1 + POCOs + ViewModel pattern + Controller with scaffolding == confusion! - asp.net-mvc-3

I'm [finally!] tackling MVC (version 3) after years of ASP.NET forms development. I have a strong background in n-tier application architecture, and I'm trying to approach this new project properly, with a clear separation of concerns, etc.
What I've done is start out with code-first by creating my POCOs. From this, the framework created my database.
Then, I implemented the Repository pattern by putting all my EF query and CRUD methods in a Repository class for each of my POCO classes in my Models assembly. This way, my Controllers don't need to know a thing about how I access my data via the EF. Great.
Finally, I created a few ViewModel classes in my Models assembly. My intent is, for certain actions (such as create and edit), I reference my ViewModel classes from the RAZOR views, instead of my POCO classes. This way, I can have my POCO class as well as a SelectList for populating a Drop Down within my ViewModel. Both populated by references to the associated Repository classes, which are called from my Controller Actions. I think I'm on a roll now:
class MyObject
{
public int ID {get;set}
[Required]
[StringLength(512)]
public string Name {get;set;}
}
class MyViewModel // ViewModel for a specific view
{
public MyObject MyModel {get;set;} // the model that is being edited
// other data the view might need, set by the controller
public string SomeMessage { get; set; }
public List<SomeObject> SomeObjects {get;set;} /// e.g. for a drop-down list
// My constructor below that populates the "SomeObjects" list, and accepts the
// "MyObject" class as a parameter in order to set the "MyModel" property above...
// ..........
}
The Problem...
Before I started using my ViewModel classes from my Controller Create and Edit actions, I passed in the POCO class directly. Everything worked fine when I hit the save button from my Edit form within my view:
Old Code:
[HttpPost]
public ActionResult Edit(MyObject mine)
{
if (ModelState.IsValid)
{
myRepository.Update(mine);
myRepository.SaveChanges();
return RedirectToAction("Index");
}
return View(mine);
}
When I hit save, my POCO class (MyObject) would be returned, automagically populated with values from the form, it would successfully save, and life was peachy.
When I switched to passing in my ViewModel (MyViewModel), everything fell apart.
I was able to refer to my ViewModel (MyViewModel) by setting the #model reference at the top of my Edit view. I was then able to populate my form fields from my POCO class (MyObject) that is part of the ViewModel. I was even able to populate the DropDownList from the SomeObjects collection in the ViewModel and preselect the correct one from my MyObject class I was editing. Everything seemed fine UNTIL...
When I hit the save button and my Controller's Edit ActionResult (POST action) was called, the MyObject class that is passed in to the ActionResult (public ActionResult Edit(MyObject mine)) was null.
Then, I tried changing the passed in object to my ViewModel (public ActionResult Edit(MyViewModel myVM)), which had the referenced MyObject class (MyModel) as null.
What am I missing?
I know it has to be something so incredibly simple that it's staring me in the face and I cannot see it!

Look at the FormCollection, the names of the keys should match the properties of the class you want to fill. This is how the default modelbinding of MVC works.

Wim,
Thanks so much for your help. I did have s parameter-less constructor, I had just omitted it from the example.
I actually tracked down the issue. In all fairness, the code I typed in was not the actual code since I didn't have it in front of me when I posted this. What the issue was is that my entity model class reference in my ViewModel actually had its set accessor as private:
public MyObject MyModel {get;private set;}
This had prevented the modelbinder from populating that property when posting back during the controller's save method.
What I'm left to do now is to move my validation logic from my EF POCO to my ViewModel, as seems to be the recommended action in this type of pattern.
Thanks for your time, and I hope this helps out other people with similar issues who are new to this framework.

Related

MVC Code First C# Retrieve Data

Ok So I am just simply needing good instructional pages on how to design a Class for retrieving data from the database.
I can find information all over on how to take an existing database and create an Entity Framework from it but I am trying to do code first.
I am able to insert Data (although I am not 100% sure how that is working) I just cannot seem to figure out how to pull the data from the database using the class(Model) that is created and display that data on a Razor page.
I have no problem with doing the studying and learning this but I am having a terrible time at finding good information that will just do a true walk through of this process.
Once again I am not looking for the Entity Framework.
Thank you for all of the help you can provide.
There is a lot of tutorials out there in the internet. Here is a small example to pull your data from table and show in the view.
Assuming you have a model class called User like this
public class User
{
public int ID { set;get;}
public string FirstName { set;get;}
}
Add properties like this to your DataContext class for each of your model entities. The property is of type DbSet.
public class YourDataContext: DbContext
{
public DbSet<User> Users {set;get;}
}
Then in your controller action method, you can create an instance of your DBContext class and access its Users property. Pass that to our view.
public class UserController : Controller
{
public ActionResult Index()
{
YourDBContext db=new YourDBContext();
var users=db.Users.ToList();
return View(users)
}
}
Have an index.cshtml view like this under Views/User folder.
#model IEnumerable User
#foreach(var user in Model)
{
<p>#user.FirstName</p>
}

MVC 3 split parameters in HttpPost action

I have an MVC 3 app and I have created a generic wrapper object, which has some navigation properties and the wrapped object of T, whose values I'm editing/displaying.
public class NavigationViewModel<T>
{
public T Model { get; set; }
public NavigationHelper NavigationHelper { get; set; }
public NavigationViewModel() { }
public NavigationViewModel(T model, NavigationHelper helper)
{
this.Model = model;
this.NavigationHelper = helper;
}
}
My controller resolves this object nicely in an action like this:
public ActionResult Foo(NavigationViewModel<Bar> viewModel)
Code in my view looks like this:
#Html.EditorFor(model => model.Model.SomeProperty)
My colleague said that that code is not nice to read. I already have a strongly typed view, the Model and this Model has another property called Model. He suggested to rename the Model property to ViewModel and I agreed with his reasoning.
Now, the code with the renamed properties does not work anymore: NavigationViewModel viewModel is null. So I changed the signature of the HttpPost method to the following and it works again:
[HttpPost]
public ActionResult Foo(NavigationHelper helper, Bar viewModel)
I like this very much! I can directly access my viewModel in code, the code in the view makes sense and the helper object does not get in the way. I haven't seen this convention before and I guess it worked before because of the naming convention. Using a property called Model hinted at how to resolve the object. Without that property, it couldn't resolve it anymore.
I would like to adopt this for other kinds of helpers that contain view-specific properties, like select-lists or other properties that I otherwise might have put in my ViewBag. Would you guys recommend this approach or will I run into trouble later on using this?
I think I have a really simple answer for you, just don't name your action parameter viewModel, so change:
public ActionResult Foo(NavigationViewModel viewModel)
public ActionResult Foo(NavigationViewModel model)
Or any other parameter name that does not collide with your ViewModel property on your NavigationViewModel class.

Constructor injection of a View Model instance used as an Action method parameter

When a view model is created you can populate the options (e.g. used in a dropdown list) into a setter property of the view model.
The problem is that when that view model is later passed as a parameter (by the framework!) into an action method, those property values has not become automagically
repopulated, so if you need to redisplay the form because of validation errors, you need to repopulate those options again.
One potential solution, which I am asking for specifically in this question, is how to make the MVC framework instantiate the view model with constructor injection, which would provide the view model constructor with an implementation of some kind of data access object (e.g. a repository) that can be used for retrieving the options when they are requested by the view (e.g. in the helper method "DropDownListFor") ?
I think the solution might have something to do with implementations of IModelBinderProvider or IModelBinder but after having experimented with these things from example code snippets here and there on the net, I am still looking for a completely working example, with downloadable executable code without any missing piece of how putting all things together.
If you are looking for some alternative discussion about how to populate a select list, e.g. with "Dependecy Lookup" instead of "Dependecy Injection" you may want to check out the following discussion:
Best way to populate SelectList for ViewModel on GET/POST
Best way to populate SelectList for ViewModel on GET/POST
Some days ago I wrote the following follow-up-question in that thread about the "Dependecy Injection" I am now looking for in this thread:
https://stackoverflow.com/a/8674525/310457
(which provides a code example about the problem I am looking for a solution of)
But instead of hoping that someone will find that old thread with a less specific title, I have created this new question with a more specific subject about what I am looking for.
And I will also provide a link from that thread into this new question for anyone that want to follow-up regarding this specific solution I am looking for.
I'm assuming you want to have your ViewModels automatically injected with something via their Constructor - for example some kind of configuration object that the View will use to determine what to show. I'm also assuming that this approach is causing a "No parameterless constructor defined for this object" error when MVC tries to automatically create and bind a model instance, from the arguments of your Controller Action. Let's also then assume that we will use a DI framework to inject the SiteConfig object into our Controllers automatically at runtime.
This means that the only problem we have to solve is how to get the injected object from our Controller into its Actions' ViewModels when they are automatically bound.
So let's define a base model for others to inherit from.
BaseViewModel
public class BaseViewModel
{
public ISiteConfig SiteConfig { get; set; }
public BaseViewModel(ISiteConfig siteConfig)
{
this.SiteConfig = siteConfig;
}
}
And now let's create a model that inherits from it.
IndexViewModel
public class IndexViewModel : BaseViewModel
{
public string SomeIndexProperty { get; set; }
public IndexViewModel (ISiteConfig siteConfig) : base(siteConfig) {}
}
And now let's define a Base Controller that our Controllers will inherit from.
BaseController
public abstract class BaseController : Controller
{
protected BaseController(ISiteConfig siteConfig)
{
_siteConfig = siteConfig;
}
private readonly ISiteConfig _siteConfig;
public ISiteConfig SiteConfig
{
get
{
return _siteConfig;
}
}
}
Now we define our actual controller.
HomeController
public HomeController: BaseController
{
public HomeController(ISiteConfig siteConfig): base(siteConfig) {}
}
Assuming we're using Ninject for DI, Ninject would be configured to automatically create the Controller and pass a concrete ISiteConfig object into its Constructor at runtime.
Now we add our Action to the Controller.
Index Action
public ActionResult Index(IndexViewModel model)
{
return View(model);
}
And so this is the point where without doing anything else, MVC will explode with a "Parameterless Constructor" error if you try to call the Index Action, because MVC can't find a ViewModel constructor that takes no arguments.
And so, the answer. We need to override the default ModelBinder.
BaseViewModelBinder
public class BaseViewModelBinder : DefaultModelBinder
{
protected override object CreateModel(ControllerContext controllerContext, ModelBindingContext bindingContext, Type modelType)
{
if (modelType == typeof(BaseViewModel) || modelType.IsSubclassOf(typeof(BaseViewModel)))
{
var baseControl = controllerContext.Controller as BaseController;
if (baseControl == null)
{
throw new Exception("The Controller must derive from BaseController");
}
var instance = Activator.CreateInstance(modelType, baseControl.SiteConfig);
bindingContext.ModelMetadata = ModelMetadataProviders.Current.GetMetadataForType(() => instance, modelType);
return instance;
}
else
{
return base.CreateModel(controllerContext, bindingContext, modelType);
}
}
}
And we need to set this as the default model binder in global.asax.cs :
protected void Application_Start()
{
...
ModelBinders.Binders.DefaultBinder = new BaseViewModelBinder();
}
That's all. As you can see, when you view the Index Action now, MVC will use our custom model binder. It will realise that the IndexViewModel derives from BaseViewModel, and so will attempt to spin up an IndexViewModel instance using the ISiteConfig it can find in the Action's Controller (because the Controller derives from BaseController).

MVC - Interface as Controller Action parameter

I want to pass one of a number of classes that implement an interface from my view back to my controller action. I use an ActionLink in my view passing the instance to my action, but it naturally fails because MVC cannot deal with interfaces via default model binding.
So :
<%=Html.ActionLink(flow.Source.Name, "Get", new {container=flow.Source})%>
is in a loop and each flow.Source conforms to IContainer.
public class Flow
{
public virtual IContainer Source { get; private set; }
}
public interface IContainer
{
//members here
}
public class File : IContainer
{}
public class Worksheet : IContainer
{}
Basically I want to call an action method :
public ActionResult Get(IContainer container)
{
// Do something
}
The reason being that I need to retrieve the state of the current container passed to my action method from the database. I use NHibernate and have entities mapped on a table per entity, so have one for File and one for Worksheet for example, so need to able to decide which data access class to use. Make sense? Probably not!
Can this be done without moving towards a base class Container? Can I stick with an interface being passed to my action method and resolve the subtype instance passed in place of the interface?
Any help with this would be gratefully appreciated.
An interface needs 'some' concrete implementation to reference when you would call your class. I think judging by your post you are aware of this : )
With that said there is 'kinda' of an approach handled here where you just create your own model binder that has to know about (or how) to map to and create a concrete type (either directly or by dependency injection)
ASP.NET MVC - Custom Model Binder on Interface Type

Who should be responsible for getting the data in MVC?

Learning ASP.NET MVC with a new project, and a little unsure of where some things should happen. I've read that ViewModels are a Good Thing (tm) and had planned on doing it in a similar fashion anyway, but I'm still not entirely clear on the responsibilities of the Model vs. the Controller.
Should the ViewModel be responsible for actually loading itself from the ORM? Thus the controller would just call ViewModel.GetObject() and pass the result back to the view?
Or should I load the data in the Controller, and then transform it into the ViewModel? Seems like that puts a lot of work in the Controller though, which is supposed to be kept somewhat lightweight.
I guess I could also have a third party that is responsible for pulling the data, then the Controller would call that and transform it for the appropriate ViewModel.
So any thoughts on what is the "best" approach?
The controller will create the viewmodel object and fill it out using the model. The model should use the ORM to get the data.
The ViewModel is always specific to the view only, and the model is specific to the domain. In CQRS you would actually just get the ViewModel and send it to the view.
From the controller you can do what ever it takes to make your CRUD happen for the view. If you use the Repo pattern thats ok, if you use NHibernate or EF directly thats cool tool. Once the ViewModel goes to the view it will be disconnected from everything like the DB, so fill it out before it gets there.
Personally I use a repository. So the controller queries a repository and gets a model, then maps the model to a view model and passes the view model to the view. Example:
public class ProductsController: Controller
{
private readonly IProductsRepository _repository;
private readonly IMapperEngine _mapper;
public ProductsController(IProductsRepository repository, IMapperEngine mapper)
{
_repository = repository;
_mapper = mapper;
}
public ActionResult Index(int id)
{
Product product = _repository.GetProduct(id);
ProductViewModel viewModel = _mapper.Map<Product, ProductViewModel>(product);
return View(viewModel);
}
}
And because this is repetitive logic I use custom action filters:
public class ProductsController: Controller
{
private readonly IProductsRepository _repository;
public ProductsController(IProductsRepository repository)
{
_repository = repository;
}
[AutoMap(typeof(Product), typeof(ProductViewModel))]
public ActionResult Index(int id)
{
Product product = _repository.GetProduct(id);
return View(product);
}
}
in this case the custom action filter intercepts the result of the action and replaces it using the corresponding mapping layer.
The way this repository is implemented is not the responsibility of the controller (whether it is an ORM, direct SQL queries, or even distant web service calls). As long as it is injected some proper implementation it will work which allows for weaker coupling between the different parts of the application and easier unit testing in isolation. So in this example it is the implementation of the repository that is responsible for fetching data.

Resources