I am going though the Apress ASP.NET MVC 3 book and trying to ensure I create Unit Tests for everything possible but after spending a good part of a day trying to work out why edit's wouldn't save (see this SO question) I wanted to create a unit test for this.
I have worked out that I need to create a unit test for the following class:
public class EFProductRepository : IProductRepository {
private EFDbContext context = new EFDbContext();
public IQueryable<Product> Products {
get { return context.Products; }
}
public void SaveProduct(Product product) {
if (product.ProductID == 0) {
context.Products.Add(product);
}
context.SaveChanges();
}
public void DeleteProduct(Product product) {
context.Products.Remove(product);
context.SaveChanges();
}
}
public class EFDbContext : DbContext {
public DbSet<Product> Products { get; set; }
}
I am using Ninject.MVC3 and Moq and have created several unit tests before (while working though the previously mentioned book) so am slowly getting my head around it. I have already (hopefully correctly) created a constructor method to enable me to pass in _context:
public class EFProductRepository : IProductRepository {
private EFDbContext _context;
// constructor
public EFProductRepository(EFDbContext context) {
_context = context;
}
public IQueryable<Product> Products {
get { return _context.Products; }
}
public void SaveProduct(Product product) {
if (product.ProductID == 0) {
_context.Products.Add(product);
} else {
_context.Entry(product).State = EntityState.Modified;
}
_context.SaveChanges();
}
public void DeleteProduct(Product product) {
_context.Products.Remove(product);
_context.SaveChanges();
}
}
BUT this is where I start to have trouble... I believe I need to create an Interface for EFDbContext (see below) so I can replace it with a mock repo for the tests BUT it is built on the class DbContext:
public class EFDbContext : DbContext {
public DbSet<Product> Products { get; set; }
}
from System.Data.Entity and I can't for the life of me work out how to create an interface for it... If I create the following interface I get errors due to lack of the method .SaveChanges() which is from the DbContext class and I can't build the interface using "DbContext" like the `EFDbContext is as it's a class not an interface...
using System;
using System.Data.Entity;
using SportsStore.Domain.Entities;
namespace SportsStore.Domain.Concrete {
interface IEFDbContext {
DbSet<Product> Products { get; set; }
}
}
The original Source can be got from the "Source Code/Downloads" on this page encase I have missed something in the above code fragments (or just ask and I will add it).
I have hit the limit of what I understand and no mater what I search for or read I can't seem to work out how I get past this. Please help!
The problem here is that you have not abstracted enough. The point of abstractions/interfaces is to define a contract that exposes behavior in a technology-agnostic way.
In other words, it is a good first step that you created an interface for the EFDbContext, but that interface is still tied to the concrete implementation - DbSet (DbSet).
The quick fix for this is to expose this property as IDbSet instead of DbSet. Ideally you expose something even more abstract like IQueryable (though this doesn't give you the Add() methods, etc.). The more abstract, the easier it is to mock.
Then, you're left with fulfilling the rest of the "contract" that you rely on - namely the SaveChanges() method.
Your updated code would look like this:
public class EFProductRepository : IProductRepository {
private IEFDbContext context;
public EFProductRepository(IEFDbContext context) {
this.context = context;
}
...
}
public interface IEFDbContext {
IDbSet<Product> Products { get; set; }
void SaveChanges();
}
BUT... the main question you have to ask is: what are you trying to test (conversely, what are you trying to mock out/avoid testing)? In other words: are you trying to validate how your application works when something is saved, or are you testing the actual saving.
If you're just testing how your application works and don't care about actually saving to the database, I'd consider mocking at a higher level - the IProductRepository. Then you're not hitting the database at all.
If you want to make sure that your objects actually get persisted to the database, then you should be hitting the DbContext and don't want to mock that part after all.
Personally, I consider both of those scenarios to be different - and equally important - and I write separate tests for each of them: one to test that my application does what it's supposed to do, and another to test that the database interaction works.
I guess your current code looks something like this (I put in the interface):
public class EFProductRepository : IProductRepository {
private IEFDbContext _context;
// constructor
public EFProductRepository(IEFDbContext context) {
_context = context;
}
public IQueryable<Product> Products {
get { return _context.Products; }
}
public void SaveProduct(Product product) {
if (product.ProductID == 0) {
_context.Products.Add(product);
} else {
_context.Entry(product).State = EntityState.Modified;
}
**_context.SaveChanges();**
}
public void DeleteProduct(Product product) {
_context.Products.Remove(product);
**_context.SaveChanges();**
}
}
public class EFDbContext : DbContext, IEFDbContext {
public DbSet<Product> Products { get; set; }
}
public interface IEFDbContext {
DbSet<Product> Products { get; set; }
}
The problem is EFProductRepository now expects an object implementing the IEFDbContext interface, but this interface does not define the SaveChanges method used at the lines I put between the asteriskes so the compiler starts complaining.
Defining the SaveChanges method on the IEFDbContext interface solves your problem:
public interface IEFDbContext {
DbSet<Product> Products { get; set; }
void SaveChanges();
}
Related
I'm working on a new project that uses CosmosDB and Entity Framework Core (via the Microsoft.EntityFrameworkCore.Cosmos NuGet package, version 5.0.7; the project itself is .NET Core 5). I'm new to both, and running into an issue I can't sort out.
In short, I need to save a complex object to the database. It's a big model that will have multiple collections of classes underneath it, each with their own properties and some with collections underneath them as well. I'm trying to configure EF with OwnsOne and OwnsMany to store these child objects underneath the top-level one. The code compiles, and will save to the database so long as all the owned objects are left empty. But whenever I put anything into an owned object, either with OwnsOne or OwnsMany, I get a pair of NullReferenceExceptions.
I've tried to strip my code down to the very basics. Here's how it currently looks.
Owner and owned classes:
public class Questionnaire
{
// Constructors
private Questionnaire() { }
public Questionnaire(Guid id)
{
Test = "Test property.";
TV = new TestQ();
Id = id;
}
public Guid Id { get; set; }
public string Test { get; set; }
public TestQ TV { get; set; }
// Public Methods
public void AddForm(Form f)
{
// not currently using this method
//Forms.Add(f);
}
}
public class TestQ
{
public TestQ()
{
TestValue = "test ownsone value";
}
public string TestValue { get; set; }
}
DbContext:
public class QuestionnaireDbContext : DbContext
{
public DbSet<Questionnaire> Questionnaires { get; set; }
public QuestionnaireDbContext(DbContextOptions<QuestionnaireDbContext> options) : base(options) { }
protected override void OnModelCreating(ModelBuilder modelBuilder)
{
modelBuilder.HasDefaultContainer(nameof(Questionnaires));
modelBuilder.Entity<Questionnaire>().HasKey(q => q.Id);
modelBuilder.Entity<Questionnaire>().OwnsOne(q => q.TV);
}
}
And the code from the service that calls the dbContext (note that this is based on a generic service that I didn't set up originally). The actual exceptions are thrown here.
public virtual TEntity Add(TEntity entity)
{
_context.Entry(entity).State = EntityState.Added;
_context.SaveChanges();
return entity;
}
Ultimately I need this to work with OwnsMany and a collection, but I figured it might be simpler to get it working with OwnsOne first. The key thing to note here is that if I comment out the line
TV = new TestQ();
in the Questionnaire class, the model persists correctly into CosmosDB. It's only when I actually instantiate an owned entity that I get the NullReferenceExceptions.
Any advice would be much appreciated! Thank you!
Well, I'm not sure why this is the case, but the issue turned out to be with how we were adding the document. Using this generic code:
public virtual async Task<TEntity> Add(TEntity entity)
{
_context.Entry(entity).State = EntityState.Added;
await _context.SaveChanges();
return entity;
}
was the issue. It works just fine if I use the actual QuestionnaireDbContext class like so:
context.Add(questionnaire);
await context.SaveChangesAsync();
Im using asp.net core. Here is the basic way to use model with controller.
public class BookController : Controller
{
private readonly ApplicationDbContext _context { get; set; }
public BookController(ApplicationDbContext context)
{
_context = context;
}
public IActionResult Create(Book model)
{
// adding new model
}
public IActionResult Edit(Book model)
{
// modifying the model
}
public IActionResult Delete(Book model)
{
// removing the model
}
}
My question: when shall/should I implement the code inside the controller? When shall/should I implement it in another class?
Something like this:
public interface IBook
{
int Add(Book book);
int Update(Book book);
int Remove(Book book);
}
public class BookData : IBook
{
private readonly ApplicationDbContext _context { get; set; }
BookData(ApplicationDbContext context)
{
_context = context
}
public int Add(Book model)
{
// ...
return _context.SaveChanges();
}
// other implements...
}
Then, calling it inside controller:
public IActionResult Create(Book model)
{
var bookData = new BookData(_context);
int result = bookData.Add(model);
// ...
}
For the interface, I think it may be useful for the case: I have many controllers that require same action/method names.
Example: MessageController requires 3 actions/methods at least (Create/Add, Edit/Update, Delete/Remove). It's same to NotificationController class, CommentController class...
So, the interface can be improved to:
public interface IMyService<T> where T : class
{
int Add(T model);
int Update(T model);
int Remove(T model);
}
public class MyService<T> : IMyService<T> where T : class
{
private readonly ApplicationDbContext _context { get; set; }
public MyService(ApplicationDbContext context)
{
_context = context;
}
public int Add(T model)
{
Type type = typeof(model);
if (type == typeof(Book))
{
// adding new book model
}
else if (type == typeof(Comment))
{
// adding new comment model
}
// ...
return -1;
}
// other implements...
}
Do I misunderstand something?
If I read it correctly with data classes you actually means repository (which is an abstraction over the persistence layer). You should always encapsulate persistence logic behind a class (be it via repository pattern, command/query pattern or request handler) and use it instead of directly using the context in your service classes.
That being said, you can directly inject your BookData to your controller instead of the ApplicationDbContext. One thing you should consider you lose in your current implementation is the Unit of Work pattern. Right now, every add will instantly persist the data.
This may not be what you want, so you should move the _context.SaveChanges(); outside of the Add/Remove/Update methods and call it explicitly. This allows you to insert i.e. 10 records and if one of them fails, nothing will be persisted to the database.
But if you call _context.SaveChanges(); after each insert and you get an error in the 8th (of 10) records, then 7 get persisted and 3 will be missing and you get inconsistent data.
Controller shouldn't contain any logic at all, only do short validation of the input model (ModelState.IsValid check) and if its okay, call the services which do all the logic and report the result back to the user. Only in very simple tutorials and guides logic is put into the controller action for reasons of simplicity. In real world applications you should never do that. Controllers are much harder to unit test than service classes.
I have some controllers (and will be more) which share some actions like those:
public ActionResult DeleteConfirmed(int id)
{
Supplier s = db.Suppliers.Find(id);
s.Deleted = true;
db.SaveChanges();
return RedirectToAction("Index");
}
public ActionResult RestoreConfirmed(int id)
{
Supplier s = db.Suppliers.Find(id);
s.Deleted = false;
db.SaveChanges();
return RedirectToAction("Index");
}
Those action are part of SuppliersController. What this does is that when I delete or restore an object, it marks the object in the database as true for deleted field (and false when it is restored).
The same behavior is shared by many other controllers like CurrenciesController, ProductsController, etc...
In the code I showed you should see that my database entity is clearly specified (Supplier) and also the repository (Suppliers).
I want to find a way to this in a generic way. I want to create a custom controller and all other controllers that shares the same behavior will extended it. In this case ProductsController will extend my DeleteRestoreController.
How can I do this in a "generic" way?
db is a DbContext
public partial class LE: DbContext
{
public LE()
: base("name=LE")
{
}
protected override void OnModelCreating(DbModelBuilder modelBuilder)
{
throw new UnintentionalCodeFirstException();
}
public virtual DbSet<Category> Categories { get; set; }
public virtual DbSet<CategoryText> CategoryTexts { get; set; }
...
}
Categories also share the same behavior.
To go one step further
public abstract class DeleteRestoreController<T> : Controller
{
public virtual Action DeleteConfirmed(int id)
{
var dbset = db.Set<T>();
var s = dbset.Find(id);
s.Deleted = true;
db.SaveChanges();
return RedirectToAction("Index");
}
}
then when defining your controller add the entity type
public class ProductsController : DeleteRestoreController<Supplier>
{
////blah
}
You can implement your DeleteRestoreController as an abstract class.
public abstract class DeleteRestoreController : Controller
{
private IRepository : Repository;
public DeleteRestoreController() { ... }
public DeleteRestoreController(IRepository Repository) { ... }
public virtual Action DeleteConfirmed(int id)
{
Supplier s = db.Suppliers.Find(id);
s.Deleted = true;
db.SaveChanges();
return RedirectToAction("Index");
}
}
If you need to differ from that behaviour in your ProductsController you can simply override that method.
public class ProductsController : DeleteRestoreController
{
public override void DeleteConfirmed()
{
//override the logic
}
}
You could always go one step further and implement a generic repository as well, but I've never gone beyond 6-8 controllers in my applications and didn't create one once.
EDIT I've just read in the comments, that the entities would change from Suppliers in the controllers, so implementing a base controller wouldn't make much sense, if you do not implement a generic interface as well. Robert Harvey has made a great point in stating the complexity has to go somewhere.
In an effort to make my entities persistent ignorant and make my repositories testable, I've implemented a repository pattern like so:
public interface IJobRepository : IRepository<Job>
{
Job GetJobById(int jobId); //Special case where I'm eager loading other entities
void SaveJob(Job job, Job originalJob);
}
public class JobRepository : IJobRepository
{
private readonly IContext _context;
public JobRepository()
{
_context = new CustomObjectContext();
}
public JobRepository(UnitOfWork unitOfWork)
{
_context = unitOfWork.Context;
}
//Basic GetAll, GetById, Add and Delete methods from IRepository<T> Interface here
//omitted for brevity
public Job GetJobById(int jobId)
{
var job = _context.Jobs.Include("Company").Include("Location").
Include("PlantInfo").Where(j => j.Jobid == jobId).SingleOrDefault();
_context.DisposeContext();
return job;
}
public void SaveJob(Job job, Job originalJob)
{
if (job.Jobid > 0)
{
// Update
_context.Jobs.Attach(originalJob);
_context.PlantInfoes.Attach(originalJob.PlantInfo);
_context.Jobs.ApplyCurrentValues(job);
_context.PlantInfoes.ApplyCurrentValues(job.PlantInfo);
Note: ApplyCurrentValues is an extension method I'm using on the ObjectSet
}
else
{
// Create
_context.Jobs.AddObject(job);
}
_context.Save();
}
}
public class UnitOfWork
{
private readonly IContext _context;
public UnitOfWork()
{
_context = new CustomObjectContext();
}
public UnitOfWork(IContext context)
{
_context = context;
}
public string Save()
{
return _context.Save();
}
internal IContext Context
{
get { return _context; }
}
}
public interface IContext
{
IObjectSet<Job> Jobs { get; }
IObjectSet<Company> Companies { get; }
IObjectSet<Location> Locations { get; }
IObjectSet<PlantInfo> PlantInfoes { get; }
string Save();
}
My ObjectContext inherits from IContext...So my understanding is that I will only use the overloaded constructor on the repository to facilitate unit tests or to use it in the case that I want to use the same context (not desirable based on this post I found on SO "Entity Framework and Connection Pooling" -- Is this right?
Also, assuming the context only gets disposed when the repository is garbage collected, I have to dispose the context explicitly to avoid the "An entity object cannot be referenced by multiple instances of IEntityChangeTracker." exception when attaching the entity prior to a save.
That said, what is the best practice for managing the DataContext in a manner that keeps your entities persistent ignorant and repositories testable?
Note: This is an asp.net webapplication; UnitOfWork and IContext implementation was based on examples from "Programming Entity Framework", Second Edition by Julia Lerman Ch24.
Thanks in advance!
Firstly, I would ensure that whatever my "consumable" object is (either repository or unit of work, depending on your setup) implements IDisposable. When your consumbed object is disposed of, then you would dispose your underlying context.
For instance, if you're using your UnitOfWork as the consumable object (the one that gets created and called in your application), it would look something like:
public class UnitOfWork : IDisposable
{
// All the other stuff you had before plus:
public void Dispose ()
{
if (_context != null)
{
_context.Dispose ();
}
}
}
(Note: This can also be done on your repositories if they're the ones being consumed directly)
And then, you have a few options in your application. If you are going to use the UnitOfWork directly, you can use it like:
public void SomeMethodThatAccessesYourData ()
{
using (var unitOfWork = new UnitOfWork (/*Load in the context*/))
{
// Access your data here.
}
}
Or, in your Web Forms or MVC object you can use constructor injection and dispose of it when the Web Forms or MVC object is disposed of:
// If you're using MVC:
public class MyController : Controller
{
private UnitOfWork _unitOfWork;
public MyController (UnitOfWork unitOfWork)
{
_unitOfWork = unitOfWork;
}
public override Dispose (bool Disposing)
{
if (Disposing && _unitOfWork != null)
{
_unitOfWork.Dispose ();
}
}
}
The same idea stands for a web forms Page.
The main reason for using the constructor overload is for Inversion of Control (IOC). It helps with both unit testing and with production code when used with an IoC Container. WebForms doesn't lend itself well to IoC, but it is really easy with MVC.
Edit
I don't really see the connection with your repositories and the unit of work. Usually you access the repositories from a unit of work or, in other implementations, you request a unit of work from your target repository. In your implementation (which I understand is not your own) there seems to be no need for both.
Edit 2
If the UoW is overkill for your application, and you know you can use IoC to inject your IContext, and you don't have very many repositories, you can do something like:
public IRepository<T> : IDisposable { }
public IJobRepository : IRepository<Job> { /* All the stuff you put here */ }
public JobRepository : IJobRepository
{
private IContext _context;
...
public void Dispose ()
{
if (_context != null)
{
_context.Dispose ();
}
}
public JobRepository (IContext context)
{
_context = context;
}
}
Then, how you use it depends on your specific task. I'm not a fan of this direct use of IRepository, but this answer is getting too long.
I want to mock my Repository object in such a way that it can still do actual DB retrieve operations. Only for Saving operations, I wanted to setup to return mock data since I don't want it to save into the DB.
How should I do it?
Thanks.
Maybe you should make your Save operation virtual and override it in a subclass which you use in your tests rather than using Moq?
First of all, your unit tests should never actually go out to the database (it is all right for integration tests, but that is a larger topic). What you want to do is pretty straightforward with Moq, though:
public class MyRepo
{
public virtual string Save(MyClass foo)
{
// perform save...
}
}
public class MyService
{
public MyRepo Repo { get; set; }
public string VerifyAndSave(MyClass foo)
{
// verify foo...
return new Repo.Save(foo);
}
}
public class MyClass()
{
public string SomeData { get; set; }
}
Notice the virtual modifiers on the methods--these are important for Moq to be able to stub them.
In your tests you could then do something like this:
[TestClass]
public class SomeTests
{
private Mock<MyRepo> MockRepo { get; set; }
private MyService Target { get; set; }
[TestInitialize]
public void Setup()
{
MockRepo = new Mock<MyRepo>();
Target = new MyService();
Target.Repo = MockRepo.Object;
}
[TestMethod]
public void MyTest()
{
const string expectedOutput = "SAVED";
MyClass exampleData = new MyClass();
MockRepo.Setup(x => x.Save(It.IsAny<MyClass>())).Returns(expectedOutput);
Target.VerifyAndSave(exampleData);
MockRepo.Verify(x => x.Save(It.IsAny<MyClass>()));
}
}
The chained calls of Setup and Returns in this case would guarantee that the calling method (i.e. VerifyAndSave) would see the value that you specified--"SAVED" in this case.
For more examples, take a look at the Moq quickstart docs.