Different return values the first and second time with Moq

I have a test like this:

    [TestCase("~/page/myaction")]
    public void Page_With_Custom_Action(string path) {
        // Arrange
        var pathData开发者_开发知识库 = new Mock<IPathData>();
        var pageModel = new Mock<IPageModel>();
        var repository = new Mock<IPageRepository>();
        var mapper = new Mock<IControllerMapper>();
        var container = new Mock<IContainer>();

        container.Setup(x => x.GetInstance<IPageRepository>()).Returns(repository.Object);

        repository.Setup(x => x.GetPageByUrl<IPageModel>(path)).Returns(() => pageModel.Object);

        pathData.Setup(x => x.Action).Returns("myaction");
        pathData.Setup(x => x.Controller).Returns("page");

        var resolver = new DashboardPathResolver(pathData.Object, repository.Object, mapper.Object, container.Object);

        // Act
        var data = resolver.ResolvePath(path);

        // Assert
        Assert.NotNull(data);
        Assert.AreEqual("myaction", data.Action);
        Assert.AreEqual("page", data.Controller);
    }

GetPageByUrl runs twice in my DashboardPathResolver, how can I tell Moq to return null the first time and pageModel.Object the second?

---

With the latest version of Moq(4.2.1312.1622), you can setup a sequence of events using SetupSequence. Here's an example:

_mockClient.SetupSequence(m => m.Connect(It.IsAny<String>(), It.IsAny<int>(), It.IsAny<int>()))
        .Throws(new SocketException())
        .Throws(new SocketException())
        .Returns(true)
        .Throws(new SocketException())
        .Returns(true);

Calling connect will only be successful on the third and fifth attempt otherwise an exception will be thrown.

So for your example it would just be something like:

repository.SetupSequence(x => x.GetPageByUrl<IPageModel>(virtualUrl))
.Returns(null)
.Returns(pageModel.Object);

---

The existing answers are great, but I thought I'd throw in my alternative which just uses System.Collections.Generic.Queue and doesn't require any special knowledge of the mocking framework - since I didn't have any when I wrote it! :)

var pageModel = new Mock<IPageModel>();
IPageModel pageModelNull = null;
var pageModels = new Queue<IPageModel>();
pageModels.Enqueue(pageModelNull);
pageModels.Enqueue(pageModel.Object);

Then...

repository.Setup(x => x.GetPageByUrl<IPageModel>(path)).Returns(pageModels.Dequeue);

---

Now you can use SetupSequence. See this post.

var mock = new Mock<IFoo>();
mock.SetupSequence(f => f.GetCount())
    .Returns(3)  // will be returned on 1st invocation
    .Returns(2)  // will be returned on 2nd invocation
    .Returns(1)  // will be returned on 3rd invocation
    .Returns(0)  // will be returned on 4th invocation
    .Throws(new InvalidOperationException());  // will be thrown on 5th invocation

---

You can use a callback when setting up your mock object. Take a look at the example from the Moq Wiki (https://github.com/Moq/moq4/wiki/Quickstart).

// returning different values on each invocation
var mock = new Mock<IFoo>();
var calls = 0;
mock.Setup(foo => foo.GetCountThing())
    .Returns(() => calls)
    .Callback(() => calls++);
// returns 0 on first invocation, 1 on the next, and so on
Console.WriteLine(mock.Object.GetCountThing());

Your setup might look like this:

var pageObject = pageModel.Object;
repository.Setup(x => x.GetPageByUrl<IPageModel>(path)).Returns(() => pageObject).Callback(() =>
            {
                // assign new value for second call
                pageObject = new PageModel();
            });

---

Adding a callback did not work for me, I used this approach instead http://haacked.com/archive/2009/09/29/moq-sequences.aspx and I ended up with a test like this:

    [TestCase("~/page/myaction")]
    [TestCase("~/page/myaction/")]
    public void Page_With_Custom_Action(string virtualUrl) {

        // Arrange
        var pathData = new Mock<IPathData>();
        var pageModel = new Mock<IPageModel>();
        var repository = new Mock<IPageRepository>();
        var mapper = new Mock<IControllerMapper>();
        var container = new Mock<IContainer>();

        container.Setup(x => x.GetInstance<IPageRepository>()).Returns(repository.Object);
        repository.Setup(x => x.GetPageByUrl<IPageModel>(virtualUrl)).ReturnsInOrder(null, pageModel.Object);

        pathData.Setup(x => x.Action).Returns("myaction");
        pathData.Setup(x => x.Controller).Returns("page");

        var resolver = new DashboardPathResolver(pathData.Object, repository.Object, mapper.Object, container.Object);

        // Act
        var data = resolver.ResolvePath(virtualUrl);

        // Assert
        Assert.NotNull(data);
        Assert.AreEqual("myaction", data.Action);
        Assert.AreEqual("page", data.Controller);
    }

---

The accepted answer, as well as the SetupSequence answer, handles returning constants.

Returns() has some useful overloads where you can return a value based on the parameters that were sent to the mocked method. Based on the solution given in the accepted answer, here is another extension method for those overloads.

public static class MoqExtensions
{
    public static IReturnsResult<TMock> ReturnsInOrder<TMock, TResult, T1>(this ISetup<TMock, TResult> setup, params Func<T1, TResult>[] valueFunctions)
        where TMock : class
    {
        var queue = new Queue<Func<T1, TResult>>(valueFunctions);
        return setup.Returns<T1>(arg => queue.Dequeue()(arg));
    }
}

Unfortunately, using the method requires you to specify some template parameters, but the result is still quite readable.

repository
    .Setup(x => x.GetPageByUrl<IPageModel>(path))
    .ReturnsInOrder(new Func<string, IPageModel>[]
        {
            p => null, // Here, the return value can depend on the path parameter
            p => pageModel.Object,
        });

Create overloads for the extension method with multiple parameters (T2, T3, etc) if needed.

---

Reached here for the same kind of problem with slightly different requirement.
I need to get different return values from mock based in different input values and found solution which IMO more readable as it uses Moq's declarative syntax (linq to Mocks).

public interface IDataAccess
{
   DbValue GetFromDb(int accountId);  
}

var dataAccessMock = Mock.Of<IDataAccess>
(da => da.GetFromDb(It.Is<int>(acctId => acctId == 0)) == new Account { AccountStatus = AccountStatus.None }
&& da.GetFromDb(It.Is<int>(acctId => acctId == 1)) == new DbValue { AccountStatus = AccountStatus.InActive }
&& da.GetFromDb(It.Is<int>(acctId => acctId == 2)) == new DbValue { AccountStatus = AccountStatus.Deleted });

var result1 = dataAccessMock.GetFromDb(0); // returns DbValue of "None" AccountStatus
var result2 = dataAccessMock.GetFromDb(1); // returns DbValue of "InActive"   AccountStatus
var result3 = dataAccessMock.GetFromDb(2); // returns DbValue of "Deleted" AccountStatus

---

We can simply declare one variable with int as a datatype. initialize it to zero, and then increase it's value as follows:

int firstTime = 0;
            repository.Setup(_ => _.GetPageByUrl<IPageModel>(path)).Returns(() =>
            {
                if (firstTime == 0)
                {
                    firstTime = 1;
                    return null;
                }
                else if(firstTime == 1)
                {
                    firstTime = 2;
                    return pageModel.Object;
                }
                else
                {
                    return null;
                }
            });

---

In some cases one needs to have the called function return different types of data based on conditions you cannot impose via the function itself. If the function accepts parameters then those can be used as conditionals to get different data.

In my case, I had a webapi call which I need to mock; earlier it was working out fine based on the input parameter, however one fine day, those parameters were converted to request headers. So since I couldn't provide a callback (no function parameters) so came up with another approach as follows

[Earlier one, when API had parameters]

this.mockedMasterAPICalls.Setup(m => m.GetCountries(It.Is<int>(ou => ou == 2), It.Is<int>(lan => lan == 1))).Returns(Task.FromResult(countryResponse));

[New one, when API had headers... The headers were getting injected into another dictionary of the API caller]

   this.mockedMasterAPICalls.Setup(m => m.RequestHeaders).Returns(new Dictionary<string, string>());
            this.mockedMasterAPICalls.Setup(m => m.GetCountries()).Returns(() =>
          {
              if (this.mockedMasterAPICalls.Object.RequestHeaders[GlobalConstants.HeaderOUInstance] == "2")
                  return Task.FromResult(countryResponse);
              else return Task.FromResult(new GetCountryResponse() { Countries = null });
          });

Note the use of the mocked object itself to make any decisions required