Rachel Appel

The Razor View Engine is a precise, useful, light language that enables you to create views for MVC projects in ASP.NET still while keeping a separation of concerns, ability to test, and pattern based development.  ASP.NET MVC developers looking for a more concise syntax now have another option that’s built-in (of course, there are many 3rd party view engines out there) with the language being a familiar light version of C#.

The Razor View engine is used to create WebMatrix sites or Visual Studio MVC applications.  When using ASP.NET MVC with either engine, you’ll stick to a style of development called “convention over configuration”, meaning that you’ll use certain naming conventions to name files, models, views, controllers, and other key application elements rather than storing lots of metadata about these same elements in a configuration file.  When using MVC in Visual Studio 2010, it’s is setup so that you’ll be guided to use convention over configuration, which becomes evident when exploring an ASP.NET MVC project.

Getting Started

Before you start, go and download these two things (as of this post; things are subject to change – these could end up in VS at some point):

• ASP.NET MVC 3
• Razor Syntax Highlighter Extension (not mandatory, but a huge help)

You can also find the Razor Syntax Highlighter by choosing Extensions Manager from the Visual Studio Tools menu.  Once installed, MVC 3 project templates are available from Visual Studio.  The MVC 3 Web Application template allows you to use either the Web Forms View Engine or the Razor View Engine, while MVC 2 Applications contain only the WF View Engine.

When creating a new MVC 3 project a new dialog box appears asking which application type, view engine, and testing framework you’d like to use.  You can and should add a test project so you can test your code, then actually write some tests in it.   The image below demonstrates selecting the internet application project template using Razor as the view engine, as well as the test framework.  The internet application MVC template adds in ASP.NET membership & security features to the project by creating the necessary model, view and controller for logging on and registering as a site user.

Verify that the project is created with three folders, one each for Models, Views, and Controllers which is the same folder structure as a MVC 2 site.  The project also contains auxiliary folders and files needed for the application such as the Content & App_Data folders.   Since this is an MVC application using the Razor View Engine, you will see a different file extension – .cshtml.  The .cshtml files are Razor View Pages written using the Razor View Engine.  If you’re not familiar with Razor syntax, I’ve blogged about it here, and the online documentation has more information as well.

The application template sets up some models, views and controllers to start with, and is now ready for new models, views, and controllers.  I’ve found the easiest way to work with MVC in VS is to start with the model, move to the controller, then create the view, so we’ll look at models first.

Models

A model is a representation of an underlying data store.  Models can be almost anything from any data source; EF or Linq2Sql Models, or a simple class.  The code the builds the model below consists of two classes, a ProductModel class and a Product class.  The ProductModel class returns a List<t> of Product objects in a property aptly named Products that represents one or more products in the data store .  Product objects contain ProductNumber, Name and Price properties and represent an individual product in the data store.  The model code representing these product objects is below:

using System.Collections.Generic;
namespace AdventureWorks.Models
{
    public class ProductModel
    {
        public List<Product> Products { get; set; }
        public ProductModel()
        {
            Products = new List<Product>();
            this.Products.Add(new Product("AB-00-J1" ,"Super Fast Bike" ,1000M));
            this.Products.Add(new Product("EE-9-OL"  ,"Durable Helmet"  ,123.45M));
            this.Products.Add(new
 Product("MMM99-54" ,"Soft Bike Seat"  ,34.99M));
        }
    }
 
    public class Product
    {
        public Product(string productNumber, string name, decimal price)
        {
            Name = name;
            Price = price;        
            ProductNumber=productNumber;
        }
        public string ProductNumber { get; set; }
        public string Name { get; set; }
        public decimal Price { get; set; }
        
    }
}

Models are handed over to views by controllers so the views can render the data from the model.  The model will extend business logic (of course, this is a simple example, business logic will likely be spread out into other applications & tiers). Models also provide validation information to the view via metadata & code.  You can add the model to the Models folder or you can reference an external data model library.  In this case, the model’s been added to the Models folder.

Controllers

Controllers are objects that inherit from the System.Web.Mvc.Controller class that match models with views.  During incoming requests, MVC’s routing system kicks in and determines which controller method to use based on routing code set in the global.asax.cs file.  A default route is set in the global.asax.cs file that works for most of the routes you’ll need, however, if need be then you can modify/add/delete your own routes.

Staying within convention, controller code files should go in the Controllers folder. The controller class must follow the naming convention of YourNameController, i.e., ProductsController.  The controller’s Index method, again following convention, must be named the same name as its view but without the file extension.  By following this pattern, the controller can have multiple actions that deliver data to multiple views, making code easier to write, maintain, and test.

namespace AdventureWorks.Controllers
{
    public class ProductsController : Controller
    {
        public ActionResult Index()
        {
            ProductModel p = new ProductModel();            
            return View(p.Products);
        }
    }
}

By observing the previous code sample and the conventions that it follows, it’s clear that the Index method maps to the ViewsProductsIndex.cshtml view.  The Index() method’s return type is an ActionResult from the System.Web.Mvc namespace.  There’s many different results that you can use for different scenarios such as ContentResult, FileResult, RedirectResult, ViewResult and others depending on your view’s needs

Views

Views are a mechanism to format and render data in a browser from the model, as specified by the @model attribute (or @inherits) at the top of the view.  Views in ASP.NET MVC with either the ASPX or Razor View engines use a code-based way to render the page rather than relying on declarative controls, which create tight coupling and are hard to test.  Using the MVC approach creates cleaner output in the browser as well as cleaner and more terse syntax in the views as well.  Views are meant to only render the markup, code and data; they’re not meant as a place to store business logic. The model and tiers accessed by the model, are the appropriate places to host business logic.  The view can have a small amount of UI validation logic,as a first line of defense for input sanitation, e.g., required, regex and range validators.

To add a view to your project, right click on the controller and select Add View.

Visual Studio will need some information about the view, which is supplied in the Add View dialog box:

1. Specify whether you want a strongly-typed view or not.  Creating a strongly-typed view allows you to refer to the underlying data structures in the classic object.property syntax.
2. Specify the view data class.  The dropdown list displays classes in your project that are part of the data model.
3. Specify the view engine.  ASPX and Razor are the options.
4. Specify the view content.  Choices are list, details, edit. etc…  Choosing this option creates .cshtml or .aspx view pages that do page rendering for you.

Here’s what choosing these options will produce in a Razor view page named ViewsProductsIndex.cshtml:

@model List<AdventureWorks.Models.Product>
@{
    View.Title = "Index";
    Layout = "~/Views/Shared/_Layout.cshtml";
}

<h2>Index</h2>
<p>
@Html.ActionLink("Create New", "Create")
</p>
<table>
  <tr>
    <th>Edit/Details/Del</th>
    <th>Product Number</th>
    <th>Name</th>
    <th>Price</th>
  </tr>
 
@foreach (var item in Model) {
  <tr>
    <td>
	@Html.ActionLink("Edit", "Edit", new { /* id=item.PrimaryKey */ }) |
	@Html.ActionLink("Details", "Details", new { /* id=item.PrimaryKey */ }) |
	@Html.ActionLink("Delete", "Delete", new { /* id=item.PrimaryKey */ })
    </td>
    <td>
	@item.ProductNumber
    </td>
    <td>
	@item.Name
    </td>
    <td>
	@String.Format("{0:F}", item.Price)
    </td>
</tr>
}
</table>

Since the view is strongly typed, the data structure can be accessed as it is – an object with properties.  The code above loops through the model and renders a <TR> for each entity that the model holds along with <TD> tags and the actual data in them (ProductNumber, name and Price).  The result is a simple list of the three products in model with accompanying action links.

The HTML source shows a clean, precise table rendered to the browser.

Summary

Rendering a list of products is very easy using MVC and the Razor View Engine, as is updating/manipulating the data (logically what you’d do next if following this example).   If you’re digging deeper into learning MVC w/Razor the next topics you’ll want to get into are jQuery and HTML validation helpers for client side validation, and decorating the model with data annotations (also for validation).  The MVC web site has a ton of great examples and tutorials.

Although many ASP.NET MVC samples use the ASPX/Web Forms view engine they’re easily compared & converted to razor pages – so don’t avoid them, it’s still a great way to learn both MVC w/Razor and the ASPX engine.

Action results are an important part of the ASP.NET MVC controller system, and definitely worth taking a good look at. Understanding how they work gives you many more choices in MVC and that will certainly help make your code better.

What is an ActionResult?

An ActionResult is a return type of a controller method, also called an action method, and serves as the base class for *Result classes. Action methods return models to views, file streams, redirect to other controllers, or whatever is necessary for the task at hand. The controller takes on this responsibility to connect system components, acting as a traffic cop.

There are many derived ActionResult types you can use to return results that are more specific for a particular view. You can quickly access the derived types of the ActionResult during development by hovering over an ActionResult in an action result’s method signature then expanding the type tool window, to see what *Results the ASP.NET Framework provides.To get a deeper look at what exactly an ActionResult type is and how it works, running the code with a breakpoint set at the end of an action result method will get the information we want. Inspecting the image below of the Watch window shows some of the properties that you can tap into that ActionResults return.

Notice the Model, TempData, ViewBag, and ViewData properties. Drilling into the Watch window even further exposes the Model property’s strongly typed collection of objects.

These objects contain everything you need when working in the view. Of course, the Model is the most heavily used object, containing our application’s schema and data. The TempData object holds data between multiple controllers, while ViewBag and ViewData objects holds data between controllers and/or views.

Since the Model property is the actual model the view uses to render the data, its data type should align with the type specified in the @model directive in the target view. As pointed out in the image above, the Model property contains a strongly typed list of Product objects (or whatever you’re returning), matching the view’s model.

Looking in the Watch window shows that the runtime converts the ActionResult to a ViewResult, however you’ll want to be more specific during design time. Swapping ActionResults for more specific/derived classes, for example, a ViewResult, produce the exact same object structure as the ActionResult. With this in mind, using more specific types increases code accuracy, maintainability, and even performance as the compiler has a much better idea of what should happen at compile time, rather than at the last minute at runtime, since no runtime conversion happens.

Action methods, action results & routing

Action methods and the routing system work together in every MVC application. MVC does this by using a set of conventions, defined as routing patterns in the global.asax.cs file. These patterns connect HTTP requests with controllers and action methods. As an HTTP Request enters the routing system, it determines which controller should serve data to which view (if any).

The global.asax.cs defines the default routing pattern, {controller}/{action}/{id}, as described in more detail here:

public static void RegisterRoutes(RouteCollection routes)
{
    routes.IgnoreRoute("{resource}.axd/{*pathInfo}");

    routes.MapRoute(
    "Default", // Route name
    "{controller}/{action}/{id}", // URL with parameters
    new { controller = "Home", action = "Index", id = UrlParameter.Optional } // Parameter defaults
    );
}

Picking apart each piece of the route pattern is a great way to digest what the routing system is doing…

• {controller} Controller maps a part of the URL to the name of the controller class. For example, a class named ProductsController, will map to URLs that start with /Products (after the domain of course). ASP.NET automatically drops the word “Controller” from the URL so is it’s more user friendly, e.g., /Products vs. /ProductsController.
• {action} Action maps a part of the URL to the individual methods inside controllers. Therefore the ProductsController.Edit(…) and ProductsController.Details(…) methods align with the /Products/Edit and Products/Details URLs, respectively. Action maps without any parameters, since the next route pattern, id, defines them.
• {id} Id maps to an identifier, such as a product id or a product name. These are controller action method parameters, so the value at the end of the URL is generally the value of the method’s argument. For example, /Products/Edit/2 matches up with the id argument in ProductsController.Edit(int id). The controller methods normally use this id to query for a single record with a matching id to return to the view.

Below is a table detailing the mappings between routes, controllers, action methods and their parameters.

You may have noticed that the Edit, Create, and Delete action methods come overloaded as pairs, and an [HttpPost] attribute decorates one method in each pair. Since [HttpGet] is the default, the other method doesn’t need to be marked. Additionally, all methods decorated with the [HttpPost] attribute are meant to accept HTML input fields located between the <form> tags. The collection parameter of type FormCollection^[1] contains all the input fields from the <form> of the HTTP Post.

Every method in a controller class returns an ActionResult by default; however, some actions will be more optimized, clear, or accurate in the code, if we are to specify a more specific action results, for instance, a ViewResult, rather than an ActionResult. Action result methods can also designate that its result will return something other than data for a view, e.g., a HttpNotFoundResult or an HttpStatusCodeResult, as shown in the code below:

Returns an HTTP 404 Not found error:

public HttpStatusCodeResult NotFound()
{
    var result = new HttpStatusCodeResult(404);
    return result;
}

Redirects to a separate view altogether:

[HttpPost]
public RedirectResult Create(FormCollection collection)
{
    // code to save form data ...     
    return new RedirectResult("Home/Index", false);
}

There’s many more ActionResult types to explore. Check them out and find something that meets the needs of your controller and/or view.

Summary

ActionResults are a key component to ASP.NET MVC, so understanding them is essential for MVC application development. Whether you need to make AJAX calls, redirect, return a file stream, or just return data to a view, ActionResults are right there in the center of all the……action.

^{[1] I recommend rather than using FormCollection, you should use a strongly typed parameter or multiple, differently typed parameters that match to data types in the HTML form. Look forward to a blog post soon to demonstrate this.}