Question

The question is not about MVC but more about code architecture.

I have a partial view that takes a CompanyModel

<%@ Control Language="C#" Inherits="System.Web.Mvc.ViewUserControl<CompanyModel>" %>

<%: Html.TextAreaFor(m => m.Name) %>
<%: Html.TextAreaFor(m => m.Location) %>

CompanyModel

public class CompanyModel
{
     [LocalizedDisplayName("Name", NameResourceType = typeof(Resources.Views.CompanyBackground))]
     public string Name{get;set;}
     [LocalizedDisplayName("Location", NameResourceType = typeof(Resources.Views.CompanyBackground))]
     public string Location{get;set;}

     public CompanyModel()
     {
          var info = Project.GetCompanyInfo();
     }
}

Project.GetCompanyInfo() is the DAL layer and gets data from the DB using Entity Framework

My question: what is the “nice” and correct way to fill CompanyModel?:

Solution 1

Create CompanyBag

public class CompanyBag
{
     public string Name{get;set;}
     public string Location{get;set;}
}

and Project.GetCompanyInfo() will return the CompanyBag and in the .ctor of CompanyModel I’ll fill the Model.

Solution 2

Project.GetCompanyInfo() will return CompanyModel (and thus no helper class is needed) and I’ll fill the Model in the .ctor of CompanyModel

Solution 3

maybe your solution?

UPDATE:
I have LoadController with Load action and this action is kind of a Factory. It will take list of views and will create them by reflection and then load to the page by Html.RenderPartial(viewName, viewModel). So, CompanyModel is just one example of the views and therefore I need each view to be independent.

So my question is: should GetCompanyInfo return a CompanyBag or a CompanyModel and I’ll fill the Model in the .ctor?

Thank you

Answer 1

Ok, following the question update, here’s my updated answer 🙂

I’m assuming the different Views LoadController populates take different types of ViewModel, and each ViewModel is populated from a different part of the domain model. Here’s how I’d tackle that:

Firstly, create each partial View as a strongly-typed subclass of ViewUserControl<TViewModel>, so LoadController can figure out which ViewModel type it needs for each of the Views it is told to create.

Secondly, define an IViewModelFactory<TViewModel> interface, and create one implementation for each ViewModel type you require; each implementation will know how to create and populate its ViewModel type from your domain model.

Finally, LoadController would figure out the type of ViewModel it needs for each view it’s been asked to create, and use the appropriate IViewModelFactory to create it.

To avoid unnecessary overhead, LoadController could find all the available IViewModelFactory implementations at runtime, and keep a static store of them against the type of ViewModel they create.

So to sum up: I would populate CompanyViewModel from a Company using a CompanyViewModelFactory, I would populate a UserViewModel from a User using a UserViewModelFactory, etc. 🙂