We Keep Coding

Too Many Dependency Inject

I currently have a class with around 40 dependency injection. It is a hard to maintain and unit test. I am not sure a good way around.
The code is done for any type of application process that is needed to process (New License, License Renewal, Student Registration, ...), there are around 80 different types applications and what sections are associated with each application type is determined by a database table.
I have a class with all of the possible properties, there are a several more than listed but you should get the idea. Each the properties have their own set of properties that are basic data types or object pointing to other classes.
class Application
{
[JsonProperty(PropertyName = "accounting")]
public Accounting Accounting { get; set; }
[JsonProperty(PropertyName = "application")]
public Application Application { get; set; }
[JsonProperty(PropertyName = "applicationType")]
public ApplicationType ApplicationType { get; set; }
[JsonProperty(PropertyName = "document")]
public List<Attachment> Document { get; set; }
[JsonProperty(PropertyName = "employment")]
public List<Employment> Employment { get; set; }
[JsonProperty(PropertyName = "enrollment")]
public Enrollment Enrollment { get; set; }
[JsonProperty(PropertyName = "individualAddressContact")]
public IndividualAddressContact IndividualAddressContact { get; set; }
[JsonProperty(PropertyName = "instructors")]
public List<Instructor> Instructors { get; set; }
[JsonProperty(PropertyName = "license")]
public License License { get; set; }
[JsonProperty(PropertyName = "licenseRenewal")]
public LicenseRenewal LicenseRenewal { get; set; }
[JsonProperty(PropertyName = "MilitaryService")]
public List<MilitaryService> MilitaryService { get; set; }
[JsonProperty(PropertyName = "paymentDetail")]
public PaymentDetail PaymentDetail { get; set; }
[JsonProperty(PropertyName = "photo")]
public List<Attachment> Photo { get; set; }
[JsonProperty(PropertyName = "portal")]
public Portal Portal { get; set; }
[JsonProperty(PropertyName = "section")]
public List<Section> Section { get; set; }
[JsonProperty(PropertyName = "testingCalendar")]
public TestingCalendar TestingCalendar { get; set; }
[JsonProperty(PropertyName = "testingScore")]
public List<TestingScore> TestingScore { get; set; }
[JsonProperty(PropertyName = "USCitizen")]
public USCitizen USCitizen { get; set; }
}
So this class is sent/received to an Angular 10 front end using Web API's.
When an application is requested the sections and the different properties are initiated and if the application has be started the progress will be reloaded. So it is possible some of properties will be pulled from the database and sent to the Angular app.
So I have something such as
Load(applicationTypeId, applicationId)
{
Get the sections for the application type
For each section in the sections
switch sectionid
case Documents
Load all of the documents required for the application type and get any documents uploaded
case Accounting
Load the payment details, if no payment made calculate the payment
case IndividualAddressContact
Load the person name/address/contact and set a few defaults if the person hasn't started.
.....
next
}
Save()
{
Save the application
switch current section
case Documents
Save all of the documents for the application
case Accounting
Save the payment details for the application
case IndividualAddressContact
Save the person name/address/contact for the application
.....
get the next section
Update the application current section
}
I have put all of the items in the switch into their own classes but in the end I still have 1 point for serialization/deserialization and still end up with to many dependencies injected. Creating a unit test with over 40 dependencies seems hard to maintain and given I won't know which properties will/won't used until an application is requested and loaded from database. I am unsure how to get around the switch, without at some point and time having to have all of the dependencies injected into 1 class.
I would appreciate some ideas of how to get around this.

"I currently have a class with around 40 dependency injection..." - Oh my gosh!
"It is a hard to maintain and unit test..." - I don't doubt that in the least!
SUGGESTED REFACTORING:
Create a class that manages "Applications" (e.g. "ApplicationManager").
Create an abstract class "Application".
One advantage of "abstract class" over "interface" here that you can put "common code" in the abstract base class.
Create a concrete subclass for each "Application" : public class NewLicense : Application, public class LicenseRenewal : Application, etc. etc.
... AND ...
Use DI primarily for those "services" that each concrete class needs.
I'll bet the constructors for your individual concrete classes will only need to inject three or four services ... instead of 40. Who knows - maybe your base class won't need any DI at all.
This is actually a design we're actually using in one of our production systems. It's simple; it's robust; it's flexible. It's working well for us :)

I would recommend using convention over configuration principle, with the Service Locator.
Declare something like IApplicationHandler interface in your program, e.g.
public interface IApplicationQueryHandler
{
Application Populate(Application application);
}
public interface IApplicationSaveHandler
{
Bool Save(Application application);
}
Then, write pieces of your code, with dependencies and such, e.g.
public class AccountingApplicationQueryHandler : IApplicationQueryHandler
{
public Application Populate(Application application) {
//// Load the payment details, if no payment made calculate the payment
return application;
}
}
public class AccountingApplicationSaveHandler : IApplicationSaveHandler
{
public Bool Save(Application application) {
//// Save the payment details for the application
return true; // this just flags for validation
}
}
// repeat for all other properties
Then in your controller, do something like
public class ApplicationController: Controller
{
public readonly IServiceProvider _serviceProvider;
public ApplicationController(IServiceProvider sp) {
_serviceProvider = sp;
}
public Application Load(string applicationTypeId, string applicationId)
{
var application = new Application(); // or get from db or whatever
var queryHandlers = _serviceProvider.GetServices(typeof(IApplicationQueryHandler));
foreach(var handler in queryHandlers) {
application = handler.Populate(application);
}
return application;
}
[HttpPost]
public bool Save(Application application)
{
var result = true;
var saveHandlers = _serviceProvider.GetServices(typeof(IApplicationSaveHandler));
foreach(var handler in queryHandlers) {
result = handler. Save(application);
}
return result;
}
}
You would need to register your handlers, which you can do e.g. like so:
var queryHandlers = Assembly.GetAssembly(typeof(IApplicationQueryHandler)).GetExportedTypes()
.Where(x => x.GetInterfaces().Any(y => y == typeof(IApplicationQueryHandler)));
foreach(queryHandler in queryHandlers) {
services.AddTransient(typeof(IApplicationQueryHandler), queryHandler);
}
// repeat the same for IApplicationSaveHandler
Now finally, you can write unit tests for part of the code like so
[TestClass]
public class AccountingApplicationQueryHandlerTests
{
[TestMethod]
public void TestPopulate()
{
// arrange
var application = new Application();
var handler = new AccountingApplicationQueryHandler(); // inject mocks here
// act
var result = handler.Populate(application);
// Assert
Assert.AreEqual(result. PaymentDetail, "whatever");
}
}
And you can test that your controller calls the right things by mocking IServiceProvider and injecting that with a couple of dummy handlers to confirm they are called correctly.

Following zaitsman's answer you also could create AggregatedApplicationQueryHandler and AggregatedApplicationSaveHandler and pass collection of concrete implementation of IApplicationQueryHandler and IApplicationSaveHandler to its constructor.
Then you don't need foreach loop inside controller(you loop over handlers inside aggregated handler) and always have only one handler passed to controller. Passing its by constructor parameter shouldn't be so much painful.
You also could create facade over some small services and aggregate theirs functions into one bigger facade service.

Dynamically change a type with C#

I am very new to C# and ServiceStack and I am working on a small project that consists on calling a third party API and loading the data I get back from the API into a relational database via ServiceStack's ORMLite.
The idea is to have each endpoint of the API have a reusable model that determines how it should be received in the API's response, and how it should be inserted into the database.
So I have something like the following:
[Route("/api/{ApiEndpoint}", "POST")]
public class ApiRequest : IReturn<ApiResponse>
{
public Int32 OrderId { get; set; }
public DateTime PurchaseDate { get; set; }
public String ApiEndpoint { get; set; }
}
public class ApiResponse
{
public Endpoint1[] Data { get; set; }
public String ErrorCode { get; set; }
public Int32 ErrorNumber { get; set; }
public String ErrorDesc { get; set; }
}
public class Endpoint1
{
[AutoIncrement]
public Int32 Id { get; set; }
[CustomField("DATETIME2(7)")]
public String PurchaseDate { get; set; }
[CustomField("NVARCHAR(50)")]
public String Customer { get; set; }
[CustomField("NVARCHAR(20)")]
public String PhoneNumber { get; set; }
public Int32 Amount { get; set; }
}
My first class represents the API's request with its route, the second class represents the API's response. The API's response is the same for all endpoints, but the only thing that varies is the structure of the Data field that comes back from that endpoint. I've defined the structure of one of my endpoints in my Endpoint1 class, and I am using it in my API's response class. As you can see, I am also defining a few attributes on my Endpoint1 class to help the ORM make better decisions later when inserting the data.
Ok, so the issue is that I have about 15 endpoints and I don't want to create 15 ApiResponse classes when I know the only thing that changes is that first Data field in the class.
So I made something like this:
public class DataModels
{
public Type getModel(String endpoint)
{
Dictionary<String, Type> models = new Dictionary<String, Type>();
models.Add("Endpoint1", typeof(Endpoint1));
// models.Add("Endpoint2", typeof(Endpoint2));
// models.Add("Endpoint3", typeof(Endpoint3));
// and so forth...
return models[endpoint];
}
}
I would like for getModel() to be called when the request is made so that I can pass in the ApiEndpoint field in the ApiRequest class and store the type that I want my Data field to have so that I can dynamically change it in my ApiResponse class.
In addition, there is the ORM part where I iterate over every endpoint and create a different table using the model/type of each endpoint. Something like this:
endpoints.ForEach(
(endpoint) =>
{
db.CreateTableIfNotExists<Endpoint1>();
// inserting data, doing other work etc
}
);
But again, I'd like to be able to call getModel() in here and with that define the model of the specific endpoint I am iterating on.
I've attempted calling getModel() on both places but I always get errors back like cannot use variable as a typeand others... so I am definitely doing something wrong.
Feel free to suggest a different approach to getModel(). This is just what I came up with but I might be ignoring a much simpler approach.

When I DID understand you correctly, you have different API-Calls which all return the same object. The only difference is, that the field "Data" can have different types.
Then you can simply change the type of data to object:
public object Data { get; set; }
And later simply cast this to the required object:
var data1=(Endpoint1[]) response.Data;

You're going to have a very tough time trying to dynamically create .NET types dynamically which requires advanced usage of Reflection.Emit. It's self-defeating trying to dynamically create Request DTOs with ServiceStack since the client and metadata services needs the concrete Types to be able to call the Service with a Typed API.
I can't really follow your example but my initial approach would be whether you can use a single Service (i.e. instead of trying to dynamically create multiple of them). Likewise with OrmLite if the Schema of the POCOs is the same, it sounds like you would be able to flatten your DataModel and use a single database table.
AutoQuery is an example of a feature which dynamically creates Service Implementations from just a concrete Request DTO, which is effectively the minimum Type you need.
So whilst it's highly recommended to have explict DTOs for each Service you can use inheritance to reuse the common properties, e.g:
[Route("/api/{ApiEndpoint}/1", "POST")]
public ApiRequest1 : ApiRequestBase<Endpoint1> {}
[Route("/api/{ApiEndpoint}/2", "POST")]
public ApiRequest2 : ApiRequestBase<Endpoint1> {}
public abstract class ApiRequestBase<T> : IReturn<ApiResponse<T>>
{
public int OrderId { get; set; }
public DateTime PurchaseDate { get; set; }
public string ApiEndpoint { get; set; }
}
And your Services can return the same generic Response DTO:
public class ApiResponse<T>
{
public T[] Data { get; set; }
public String ErrorCode { get; set; }
public Int32 ErrorNumber { get; set; }
public String ErrorDesc { get; set; }
}
I can't really understand the purpose of what you're trying to do so the API design is going to need modifications to suit your use-case.
You're going to have similar issues with OrmLite which is a Typed code-first POCO ORM where you're going to run into friction trying to use dynamic types which don't exist at Runtime where you'll likely have an easier time executing Dynamic SQL since it's far easier to generate a string than a .NET Type.
With that said GenericTableExpressions.cs shows an example of changing the Table Name that OrmLite saves a POCO to at runtime:
const string tableName = "Entity1";
using (var db = OpenDbConnection())
{
db.DropAndCreateTable<GenericEntity>(tableName);
db.Insert(tableName, new GenericEntity { Id = 1, ColumnA = "A" });
var rows = db.Select(tableName, db.From<GenericEntity>()
.Where(x => x.ColumnA == "A"));
Assert.That(rows.Count, Is.EqualTo(1));
db.Update(tableName, new GenericEntity { ColumnA = "B" },
where: q => q.ColumnA == "A");
rows = db.Select(tableName, db.From<GenericEntity>()
.Where(x => x.ColumnA == "B"));
Assert.That(rows.Count, Is.EqualTo(1));
}
Which uses these extension methods:
public static class GenericTableExtensions
{
static object ExecWithAlias<T>(string table, Func<object> fn)
{
var modelDef = typeof(T).GetModelMetadata();
lock (modelDef)
{
var hold = modelDef.Alias;
try
{
modelDef.Alias = table;
return fn();
}
finally
{
modelDef.Alias = hold;
}
}
}
public static void DropAndCreateTable<T>(this IDbConnection db, string table)
{
ExecWithAlias<T>(table, () => {
db.DropAndCreateTable<T>();
return null;
});
}
public static long Insert<T>(this IDbConnection db, string table, T obj, bool selectIdentity = false)
{
return (long)ExecWithAlias<T>(table, () => db.Insert(obj, selectIdentity));
}
public static List<T> Select<T>(this IDbConnection db, string table, SqlExpression<T> expression)
{
return (List<T>)ExecWithAlias<T>(table, () => db.Select(expression));
}
public static int Update<T>(this IDbConnection db, string table, T item, Expression<Func<T, bool>> where)
{
return (int)ExecWithAlias<T>(table, () => db.Update(item, where));
}
}
But it's not an approach I'd take personally, if I absolutely needed (and I'm struggling to think of a valid use-case outside of table-based Multitenancy or sharding) to save the same schema in multiple tables I'd just be using inheritance again, e.g:
public class Table1 : TableBase {}
public class Table2 : TableBase {}
public class Table3 : TableBase {}

Inject a property (recursively) without passing it as a parameter

I'm trying to achieve maybe something that might be impossible.
We have a big MVC 5 application. I created a small MVC project to simulate and explain what I want to apply into that big MVC project.
I have a controller that has unique Id. In this sample project the unique Id is regenerated for each request. In the MVC project, it is a bit more complex and different. However it's not relevant in the scope of this example.
public class FooController : Controller
{
public string UniqueId = Guid.NewGuid().ToString("N");
public ActionResult Index()
{
var worker = new WorkerA();
worker.DoWork();
return View();
}
}
The FooController creates WorkerA which creates WorkerB which creates WorkerC and so on. The workers are not the same. They don't have the same interface/implementation. To make the example simple I made them look similar.
Here's the Workers:
public class WorkerA
{
public string UniqueId = string.Empty;
public void DoWork()
{
var worker = new WorkerB();
worker.DoWork();
//...
//...
}
}
public class WorkerB
{
public string UniqueId = string.Empty;
public void DoWork()
{
var worker = new WorkerC();
worker.DoWork();
}
}
I want to have inject the property UniqueId into the worker without having to passing it as a parameter.
I want to avoid having to do this:
public WorkerA(string uniqueId)
{
UniqueId = uniqueId;
}
But I need to do the same for all the other workers.
EDIT
Is there a way to acheive that with ninject?

You can achieve what you want using Microsoft.Practices.Unity in the following manner:
public class WorkerA
{
[Dependency]
public string UniqueId { get; set; }
}
public class WorkerB
{
[Dependency]
public string UniqueId { get; set; }
}
And after that :
var container = new UnityContainer();
container.RegisterType<WorkerA>(new InjectionProperty(nameof(WorkerA.UniqueId),"WorkerAValue"));
container.RegisterType<WorkerA>(new InjectionProperty(nameof(WorkerB.UniqueId), "WorkerBValue"));
Later, you can request the instances from the container with the desired properties configured:
var workerA = container.Resolve<WorkerA>();
var workerB = container.Resolve<WorkerB>();

You can do something like:
worker.GetType().GetField("prop")?.SetValue(worker, "guid");

You could create a singleton class to manage the GUID and deliver it to the child classes that way. This way you can still do it in a constructor but not have to pass it as a parameter
public class GUIDManager
{
private static GUIDManager _instance;
private Guid _activeGuid;
public Guid ActiveGuid {
get { return _activeGuid; }
set { _activeGuid = value; }
}
private GUIDManager()
{
if (_activeGuid == null)
_activeGuid = new Guid();
}
public static GUIDManager GetInstance()
{
if(_instance == null)
{
_instance = new GUIDManager();
}
return _instance;
}
}
public class WorkerB
{
public string UniqueId = string.Empty;
public WorkerB()
{
var manager = GUIDManager.GetInstance();
UniqueId = manager.ActiveGuid.ToString();
}
public void DoWork()
{
var worker = new WorkerC();
worker.DoWork();
}
}

From your question i'm not entirely clear about all the workers in the same request getting the same ID or not. If they all should get the same ID then it's simple:
Wrap the ID in a class and use InRequestScope():
public class BrowserTabId
{
public string browserTabId;
public BrowserTabId(string tabId)
{
if(string.IsNullOrEmpty(tabId))
{
throw new NullArgumentException();
}
this.browserTabId = tabId;
}
public string Id { get { return this.browserTabId; } }
}
Bind<BrowserTabId>()
.ToMethod(ctx =>
new BrowserTabId(HttpContext.Items["BrowserTabId"] as string)))
.InRequestScope();
For testability reasons you can also slap on an interface IUniqueRequestId and create the binding for that.
This will result in all workers / objects created during the same request receiveing the same BrowserTabId. If you don't want to use c-tor injection you can use property injection instead. If you don't want to inject the value all the type, then use a When(..) condition to specify when to inject and when not to. Combine this with the null-object pattern to keep ninject from complaining that it can't inject a requested type.
Property Injection
Adapt a worker as follows:
public class WorkerA
{
[Inject]
public BrowserTabId BrowserTabId { get; set; }
....
}
Note, however, for this to work, like normal constructor injection, it is necessary that either the WorkerA is instanciated by ninject or that Ninject is informed about its existence by Ninject.Inject(workerAInstance)
Scoping
Since you mention that the lifetime of the ID in your actual application is somewhat more complicated, I guess you will have to use something different than InRequestScope - maybe roll your own scope (by using InScope(...)). Or Maybe, InCallScope() is as viable alternative. However, without knowing what exactly it is what you need, it's a bit difficult to advise you properly.

Saving an Item in EF with existing childrens

Im having some problems saving an object (FeatureType) that have a 1-M relationship with Section.
public class FeatureType
{
public int Id { get; set; }
public string Name { get; set; }
[ForeignKey("SectionId")]
public Section Section { get; set; }
public virtual List<ItemType> ItemTypes { set; get; }
}
public class Section
{
public int Id { get; set; }
public string Name { get; set; }
public int Order { get; set; }
public virtual List<FeatureType> Features { get; set; }
}
If The ItemTypes are new i have no problem and the insert is done correctly.
But if i want to add some existing ItemTypes im getting this Error:
An entity object cannot be referenced by multiple instances of
IEntityChangeTracker.
I have been reading about this problem but i havent found a way to solve it, and it might be because of how its designed my application.
Whem im mappinig from my viewModel to my Model, im getting the section ID and getting the section Object from my SectionRepository as this:
private Section GetSection()
{
var section = _sectionRepository.GetSection(SectionId);
return section;
}
And this is what is giving me the problem, as the section is now been tracked by the SectionRepository that have its own context.
How can i solve this? I have tried just creating a new section with the existing ID but it just create me an empty object.
private Section GetSection()
{
var section = new Section{Id=SectionId};
return section;
}
UPDATE
To save my entity i just use :
_repository.Create(featureType.ToModel());
public FeatureType ToModel()
{
var ft = new FeatureType
{
Name = Name,
ControlType = (ControlType)ControlType,
Order = Order,
Required = Required,
RequiredText = RequiredText,
ItemTypes = GetItemTypes().ToList(),
Section = GetSection(),
};
return ft;
}
UPDATE 2: This is how i have my repositories, i wouldn't like to manage any EF in my controller but with some kind of repository or service.
public class EFBaseRepository
{
protected MyContext Db = new MyContext();
public void Dispose(bool disposing)
{
Db.Dispose();
}
}
public class EFFeatureTypeRepository : EFBaseRepository, IFeatureTypeRepository
{
public IQueryable<FeatureType> GetFeatureTypes
{
get { return Db.FeatureTypes.Include("Section").Include("ItemTypes"); }
}
public Message Create(FeatureType feature)
{
try
{
Db.FeatureTypes.Add(feature);
Db.SaveChanges();
return new Message();
}
catch (Exception e)
{
throw;
// return new Message(e, string.Format("Error Creating {0}", feature.GetType()));
}
}
//..Other Methods
}

You say that the SectionRepository has its own context. That is going to cause you problems. The repositories should share a context. The context is a combination of the unit of work and repository patterns. You need to separate the two patterns:
How to migrate towards unit-of-work and repository pattern
EDIT
You can avoid having the DbContext in the Controller by implementing your own Unit Of Work pattern.
public interface IUnitOfWork : IDisposable
{
ISectionRepository SectionRepository {get;}
//etc
int Save();
}
then in your controller:
public ActionResult Create(FeatureTypeCreate featureType)
{
_Uow.SectionRepository.Create(featureType.ToModel());
_Uow.Save(); //Saving is the responsibility of the Unit Of Work
//not the Repository
}
More references:
Implementing the Repository and Unit of Work
Repository and Unit of Work in Entity Framework
John Papa's original source code

Simply, the error you're getting means that the entities were returned from a different instance of your DbContext than from which they are now trying to be saved. Make sure that you're not doing something like using two different usings around your repository and that your repository always makes use of the same DbContext per instantiation.

DDD Approach to Access External Information

I have an existing bank application classes as shown below. The banks account can be of SavingsBankAccount or FixedBankAccount. There is an operation called IssueLumpSumInterest. For FixedBankAccount, the balance need to be updated only if the owner of the account has no other account.
This demands the FixedBankAccount object to know about other accounts of the account owner. How to do this by following SOLID/DDD/GRASP/Information Expert pattern?
namespace ApplicationServiceForBank
{
public class BankAccountService
{
RepositoryLayer.IRepository<RepositoryLayer.BankAccount> accountRepository;
ApplicationServiceForBank.IBankAccountFactory bankFactory;
public BankAccountService(RepositoryLayer.IRepository<RepositoryLayer.BankAccount> repo, IBankAccountFactory bankFact)
{
accountRepository = repo;
bankFactory = bankFact;
}
public void IssueLumpSumInterest(int acccountID)
{
RepositoryLayer.BankAccount oneOfRepositroyAccounts = accountRepository.FindByID(p => p.BankAccountID == acccountID);
int ownerID = (int) oneOfRepositroyAccounts.AccountOwnerID;
IEnumerable<RepositoryLayer.BankAccount> accountsForUser = accountRepository.FindAll(p => p.BankUser.UserID == ownerID);
DomainObjectsForBank.IBankAccount domainBankAccountObj = bankFactory.CreateAccount(oneOfRepositroyAccounts);
if (domainBankAccountObj != null)
{
domainBankAccountObj.BankAccountID = oneOfRepositroyAccounts.BankAccountID;
domainBankAccountObj.AddInterest();
this.accountRepository.UpdateChangesByAttach(oneOfRepositroyAccounts);
//oneOfRepositroyAccounts.Balance = domainBankAccountObj.Balance;
this.accountRepository.SubmitChanges();
}
}
}
public interface IBankAccountFactory
{
DomainObjectsForBank.IBankAccount CreateAccount(RepositoryLayer.BankAccount repositroyAccount);
}
public class MySimpleBankAccountFactory : IBankAccountFactory
{
public DomainObjectsForBank.IBankAccount CreateAccount(RepositoryLayer.BankAccount repositroyAccount)
{
DomainObjectsForBank.IBankAccount acc = null;
if (String.Equals(repositroyAccount.AccountType, "Fixed"))
{
acc = new DomainObjectsForBank.FixedBankAccount();
}
if (String.Equals(repositroyAccount.AccountType, "Savings"))
{
//acc = new DomainObjectsForBank.SavingsBankAccount();
}
return acc;
}
}
}
namespace DomainObjectsForBank
{
public interface IBankAccount
{
int BankAccountID { get; set; }
double Balance { get; set; }
string AccountStatus { get; set; }
void FreezeAccount();
void AddInterest();
}
public class FixedBankAccount : IBankAccount
{
public int BankAccountID { get; set; }
public string AccountStatus { get; set; }
public double Balance { get; set; }
public void FreezeAccount()
{
AccountStatus = "Frozen";
}
public void AddInterest()
{
//TO DO: Balance need to be updated only if the person has no other accounts.
Balance = Balance + (Balance * 0.1);
}
}
}
READING
Issue in using Composition for “is – a “ relationship
Implementing Business Logic (LINQ to SQL)
http://msdn.microsoft.com/en-us/library/bb882671.aspx
Architecting LINQ to SQL applications
Exploring N-Tier Architecture with LINQ to SQL
http://randolphcabral.wordpress.com/2008/05/08/exploring-n-tier-architecture-with-linq-to-sql-part-3-of-n/
Confusion between DTOs (linq2sql) and Class objects!
Domain Driven Design (Linq to SQL) - How do you delete parts of an aggregate?

The first thing I noticed was the improper use of the bank account factory. The factory, pretty much as you have it, should be used by the repository to create the instance based on the data retrieved from the data store. As such, your call to accountRepository.FindByID will return either a FixedBankAccount or SavingsBankAccount object depending on the AccountType returned from the data store.
If the interest only applies to FixedBankAccount instances, then you can perform a type check to ensure you are working with the correct account type.
public void IssueLumpSumInterest(int accountId)
{
var account = _accountRepository.FindById(accountId) as FixedBankAccount;
if (account == null)
{
throw new InvalidOperationException("Cannot add interest to Savings account.");
}
var ownerId = account.OwnerId;
if (_accountRepository.Any(a => (a.BankUser.UserId == ownerId) && (a.AccountId != accountId)))
{
throw new InvalidOperationException("Cannot add interest when user own multiple accounts.");
}
account.AddInterest();
// Persist the changes
}
NOTE: FindById should only accept the ID parameter and not a lambda/Func. You've indicated by the name "FindById" how the search will be performed. The fact that the 'accountId' value is compared to the BankAccountId property is an implementation detail hidden within the method. Name the method "FindBy" if you want a generic approach that uses a lambda.
I would also NOT put AddInterest on the IBankAccount interface if all implementations do not support that behavior. Consider a separate IInterestEarningBankAccount interface that exposes the AddInterest method. I would also consider using that interface instead of FixedBankAccount in the above code to make the code easier to maintain and extend should you add another account type in the future that supports this behavior.

From reading your requirement, here is how I would do it:
//Application Service - consumed by UI
public class AccountService : IAccountService
{
private readonly IAccountRepository _accountRepository;
private readonly ICustomerRepository _customerRepository;
public ApplicationService(IAccountRepository accountRepository, ICustomerRepository customerRepository)
{
_accountRepository = accountRepository;
_customerRepository = customerRepository;
}
public void IssueLumpSumInterestToAccount(Guid accountId)
{
using (IUnitOfWork unitOfWork = UnitOfWorkFactory.Create())
{
Account account = _accountRepository.GetById(accountId);
Customer customer = _customerRepository.GetById(account.CustomerId);
account.IssueLumpSumOfInterest(customer);
_accountRepository.Save(account);
}
}
}
public class Customer
{
private List<Guid> _accountIds;
public IEnumerable<Guid> AccountIds
{
get { return _accountIds.AsReadOnly();}
}
}
public abstract class Account
{
public abstract void IssueLumpSumOfInterest(Customer customer);
}
public class FixedAccount : Account
{
public override void IssueLumpSumOfInterest(Customer customer)
{
if (customer.AccountIds.Any(id => id != this._accountId))
throw new Exception("Lump Sum cannot be issued to fixed accounts where the customer has other accounts");
//Code to issue interest here
}
}
public class SavingsAccount : Account
{
public override void IssueLumpSumOfInterest(Customer customer)
{
//Code to issue interest here
}
}
The IssueLumpSumOfInterest method on the Account aggregate requires the Customer aggregate to help decide whether interest should be issued.
The customer aggregate contains a list of account IDs - NOT a list of account aggregates.
The base class 'Account' has a polymorphic method - the FixedAccount checks that the customer doesn't have any other accounts - the SavingsAccount doesn't do this check.

2 min scan answer..
Not sure why there is a need for 2 representations of a BankAccount
RepositoryLayer.BankAccount and DomainObjectsForBank.IBankAccount. Hide the persistence layer coupled one.. deal with just the domain object in the service.
Do not pass/return Nulls - I think is good advice.
The finder methods look like the LINQ methods which select items from a list of collection. Your methods look like they want to get the first match and exit..in which case your parameters can be simple primitives (Ids) vs lambdas.
The general idea seems right. The service encapsulates the logic for this transaction - not the domain objects. If this changes, only one place to update.
public void IssueLumpSumInterest(int acccountID)
{
var customerId = accountRepository.GetAccount(accountId).CustomerId;
var accounts = accountRepository.GetAccountsForCustomer(customerId);
if ((accounts.First() is FixedAccount) && accounts.Count() == 1)
{
// update interest
}
}

Things that strike me as weird:
Your IBankAccount has a method FreezeAccount, but I presume that all accounts would have quite similar behavior? Perhaps a BankAccount class is warranted that implements some of the interface?
AccountStatus should probably be an enum? What should happen if an account is "Forzen"?