Currently we have implemented a repository pattern at work. All our repositories sit behind their own interfaces and are mapped via Ninject. Our project is quite large and there are a couple quirks with this pattern I'm trying to solve.
First, there are some controllers where we need upwards of 10 to 15 repositories all in the same controller. The constructor gets rather ugly when asking for so many repositories. The second quirk reveals itself after you call methods on multiple repositories. After doing work with multiple repositories we need to call the SaveChanges method, but which repository should we call it on? Every repository has one. All repositories have the same instance of the Entity Framework data context injected so picking any random repository to call save on will work. It just seems so messy.
I looked up the "Unit Of Work" pattern and came up with a solution that I think solves both problems, but I'm not 100% confident in this solution. I created a class called DataBucket.
// Slimmed down for readability
public class DataBucket
{
private DataContext _dataContext;
public IReportsRepository ReportRepository { get; set; }
public IEmployeeRepository EmployeeRepository { get; set; }
public IDashboardRepository DashboardRepository { get; set; }
public DataBucket(DataContext dataContext,
IReportsRepository reportsRepository,
IEmployeeRepository employeeRepository,
IDashboardRepository dashboardRepository)
{
_dataContext = dataContext;
this.ReportRepository = reportsRepository;
this.EmployeeRepository = employeeRepository;
this.DashboardRepository = dashboardRepository;
}
public void SaveChanges()
{
_dataContext.SaveChanges();
}
}
This appears to solve both issues. There is now only one SaveChanges method on the data bucket itself and you only inject one object, the data bucket. You then access all the repositories as properties. The data bucket would be a little messy looking since it would be accepting ALL (easily 50 or more) of our repositories in its constructor.
The process of adding a new repository would now include: creating the interface, creating the repository, mapping the interface and repository in Ninject, and adding a property to the data bucket and populating it.
I did think of an alternative to this that would eliminate a step from above.
public class DataBucket
{
private DataContext _dataContext;
public IReportsRepository ReportRepository { get; set; }
public IEmployeeRepository EmployeeRepository { get; set; }
public IDashboardRepository DashboardRepository { get; set; }
public DataBucket(DataContext dataContext)
{
_dataContext = dataContext;
this.ReportRepository = new ReportsRepository(dataContext);
this.EmployeeRepository = new EmployeeRepository(dataContext);
this.DashboardRepository = new DashboardRepository(dataContext);
}
public void SaveChanges()
{
_dataContext.SaveChanges();
}
}
This one pretty much eliminates all the repository mappings in Ninject because they are all instantiated in the data bucket. So now the steps to adding a new repository include: Create interface, create repository, add property to data bucket and instantiate.
Can you see any flaws with this model? On the surface it seems much more convenient to consume our repositories in this way. Is this a problem that has been addressed before? If so, what is the most common and/or most efficient approach to this issue?
First, there are some controllers where we need upwards of 10 to 15 repositories all in the same controller.
Say hello to Abstract factory pattern. Instead of registering all repositories in Ninject and injecting them to controllers register just single implementation of the factory which will be able to provide any repository you need - you can even create them lazily only if the controller really needs them. Than inject the factory to controller.
Yes it also has some disadvantages - you are giving controller permission to get any repository. Is it problem for you? You can always create multiple factories for some sub systems if you need or simply expose multiple factory interfaces on single implementation. It still doesn't cover all cases but it is better than passing 15 parameters to constructor. Btw. are you sure those controllers should not be split?
Note: This is not Service provider anti-pattern.
After doing work with multiple repositories we need to call the SaveChanges method, but which repository should we call it on?
Say hello to Unit of Work pattern. Unit of Work is logical transaction in your application. It persists all changes from logical transaction together. Repository should not be responsible for persisting changes - the unit of work should be. Somebody mentioned that DbContext is implementation of Repository pattern. It is not. It is implementation of Unit of Work pattern and DbSet is implementation of Repository pattern.
What you need is central class holding the instance of the context. The context will be also passed to repositories because they need it to retrieve data but only the central class (unit of work) will offer saving changes. It can also handle database transaction if you for example need to change isolation level.
Where should be unit of work handled? That depends where your logical operation is orchestrated. If the operation is orchestrated directly in controller's actions you need to have unit of work in the action as well and call SaveChanges once all modifications are done.
If you don't care about separation of concerns too much you can even combine unit of work and factory into single class. That brings us to your DataBucket.
I think you are absolutely right to use the Unit of Work pattern in this case. Not only does this prevent you from needing a SaveChanges method on every repository, it provides you a nice way to handle transactions from within code rather than in your database itself. I included a Rollback method with my UOW so that if there was an exception I could undo any of the changes the operation had already made on my DataContext.
One thing you could do to prevent weird dependency issues would be to group related repositories on their own Unit of Work, rather than having one big DataBucket that holds every Repository you have (if that was your intent). Each UOW would only need to be accessible at the same level as the repositories it contained, and other repositories should probably not depend on other UOWs themselves (your repositories shouldn't need to use other repositories).
If wanted to be an even bigger purist of the pattern, you could also structure your UOWs to represent just that, a single Unit of Work. You define them to represent a specific operation in your domain, and provide it with the repositories required to complete that operation. Individual repositories could exist on more than one UOW, if it made sense to be used by more than one operation in your domain.
For example, a PlaceCustomerOrderUnitOfWork may need a CustomerRepository, OrderRepository, BillingRepository, and a ShippingRepository
An CreateCustomerUnitOfWork may need just a CustomerRepository. Either way, you can easily pass that dependency around to its consumers, more fine grained interfaces for your UOW can help target your testing and reduce the effort to create a mock.
The notion of every repository having a SaveChanges is flawed because calling it saves everything. It is not possible to modify part of a DataContext, you always save everything. So a central DataContext holder class is a good idea.
Alternatively, you could have a repository with generic methods that can operate on any entity type (GetTable<T>, Query<T>, ...). That would get rid of all those classes and merge them into one (basically, only DataBucket remains).
It might even be the case that you don't need repositories at all: You can inject the DataContext itself! The DataContext by itself is a repository and a full fledged data access layer. It doesn't lend itself to mocking though.
If you can do this depends on what you need the "repository" do provide.
The only issue with having that DataBucket class would be that this class needs to know about all entities and all repositories. So it sits very high in the software stack (at the top). At the same time it is being used by basically everything so it sits at the bottom, too. Wait! That is a dependency cycle over the whole codebase.
This means that everything using it and everything being used by it must sit in the same assembly.
What I have done in the past was to create child injection containers (I was using Unity) and register a data context with a ContainerControlledLifetime. So that when the repositories are instantiated, they always have the same data context injected into them. I then hang on to that data context and when my "Unit of Work" is complete, I call DataContext.SaveChanges() flushing all the changes out to the database.
This has some other advantages such as (with EF) some local caching, such that if more than one repository needs to get the same entity, only the first repository actually causes a database round trip.
It's also a nice way to "batch up" the changes and make sure they execute as a single atomic transaction.
Related
Overview
I am making a .net-core web API with simple CRUD operations.
I have made the GET methods and got them up and running, however when I try to implement the Create method using dbContext.Add(myItem) I am not able to use dbContext.SaveChanges() afterwards.
My current work is available at:
https://github.com/petermefrandsen/TKD-theory-API
So far
I have tried adding a overwriting method to my database context.
Additionally I have tried adding the entity framework reference to the project.
As I do use interfaces for loose coupling I am at a loss when I comes to comparing with tutorials and other peoples similar problems. (read I am fairly new to c#).
Code
Controller:
[Route("dan/")]
[HttpPost]
public ActionResult<DanTheoryItem> PostDanTheoryItem(DanTheoryItem danTheoryItem)
{
_context.PostDanTheoryItem(danTheoryItem);
return new ActionResult<DanTheoryItem>(danTheoryItem);
}
IContext:
DanTheoryItem PostDanTheoryItem(DanTheoryItem danTheoryItem);
Context:
public DanTheoryItem PostDanTheoryItem(DanTheoryItem danTheoryItem)
{
var theoryItem = new DbDanTheoryItems
{
Id = danTheoryItem.Id,
KoreanTheoryItemId = danTheoryItem.KoreanTheoryItemId,
MainCategory = danTheoryItem.MainCategory,
SubCategory = danTheoryItem.SubCategory,
SubToSubCategory = danTheoryItem.SubToSubCategory,
NameLatin = danTheoryItem.NameLatin,
NamePhonetic = danTheoryItem.NamePhonetic,
NameAudio = danTheoryItem.NameAudio
};
_dbContext.DanTheoryItems.Add(theoryItem);
//_dbContext.SaveChanges();
return danTheoryItem;
}
Desired result
I'd like to have the controller call the context methods that will write the desired data to the database.
Your interface doesn't contain a SaveChanges method. Since you are using dependency injection only the methods in your interface will be available to your controller.
If you inherit from System.Data.Entity.IDbContext class in your custom interface, the method will be exposed to you in your controller.
Instantiating an instance of your DbLokisaurTKDTheoryAppContext class in your controller will expose the SaveChanges method as well.
It's hard to say exactly because you've neglected to post certain key portions of your code. However, my best guess is that you "context" is being provided IDbContext (of your own creating), instead of DbContext, and your interface doesn't define a SaveChanges method.
Honestly, just get rid of IDbContext. The point of an interface is to define a contract between multiple implementations. Here, there can be only one implementation: DbContext, and DbContext isn't even aware of this interface. Just inject your derived DbContext directly.
Using an interface isn't a magic wand. You have a hard dependency on Entity Framework here, so interface or not, you're tightly coupled. That's not necessary a bad thing though. EF is serving as your data layer, and this application is data driven; it's going to be tightly coupled to the data layer no matter what.
The API itself serves as your abstraction. Other layers will presumably just use the API, so no further abstraction is necessary, and in fact just adds more maintenance concern with no added benefit.
I am trying to figure out how to correctly implement UoW and Repository pattern with Entity Framework. I saw a lot of post against it but it still seems like the right way to go.
I am trying to do it according to this blog post. The last thing I am trying to figure out is how to inject the Repositories into the UoW but in a way that lets me do it on demand. The number of repositories may grow and so will the constructor then. Furthermore instantiating all of the repositories for an operation that may require only 1 or 2 seems like a waste of resources.
How do I do it in a way that lets me write unit tests fairly easily?
The only way I found out, that lets me inject repositories NOT in the constructor (so they are not all instantiated, even when they are not needed for a particular operation) is by doing it in the getter:
private IGenericRepository<Blog> _blogRepository;
private IGenericRepository<Post> _postRepository;
public UnitOfWork(BloggingContext bloggingContext)
{
_bloggingContext = bloggingContext;
}
public IGenericRepository<Blog> BlogRepository
{
get
{
return _blogRepository = _blogRepository ?? new GenericRepository<Blog>(_bloggingContext);
}
}
However, this approach generates a lot of noise in the code because when I will have 50 repositories I will need 50 props.
You might want to combine the described approach (which I like) with the generic repository approach described here https://cpratt.co/truly-generic-repository/
Given proper implementation of the UoW pattern from the blog post you reference, you won't need the anything but the IReadOnlyRepository from the article - it will provide everything you need for the repository to be.
I'm refactoring an existing MVC.Net application to include the unit of work pattern to make data management a bit more obvious and straight forward.
The application is currently split into
Presentation/UI (MVC Controllers delivering views OR JsonResults for AngularJS)
Business Logic (Containing well... business logic)
DAL (Repositories and EF)
I'm having a hard time trying to figure out how I need to be structuring dependency injection and UoW passing to keep things sensible and testable.
I'm anticipating something like the following to be an example:
public class SomeMVCController : Controller
{
private readonly IStoreFrontLogic _storeFrontLogic;
public SomeMVCController(IStoreFrontLogic storeFrontLogic)
{
_storeFrontLogic = storeFrontLogic;
var uow = new UnitOfWork(User);
_storeFrontLogic.UnitOfWork = uow;
}
public ActionResult SomeRequest()
{
var myViewModel = _storeFrontLogic.OffersForUser();
return View(myViewModel);
}
}
public class StoreFrontLogic : IStoreFrontLogic
{
public UnitOfWork unitOfWork;
public OffersModel OffersForUser()
{
//some logic taking into account the current user in the uow
var prevOrders = unitOfWork.OrdersRepo.GetUsersOrders();
// special offers logic
return specialOffers;
}
}
Does this seem sensible?
I'm not too keen on the requirement to manually push the uow into my logic classes whenever they're required. Is there a more sensible way?
As I said above, this is hard to answer without a specific question or specific domain model but I'll give it a shot.
My understanding of such things is focused pretty heavily through a Domain Driven Design lens.
First of, you should read this series of papers on effective aggregate design. The fact that you need units of work and to do queries from inside your domain classes implies that your model needs work.
Some other thoughts on UOW - having uow produce your repositories is good, but I think you will likely start hitting lots of difficulties with implementation. UoW is super useful in small targeted areas but is very difficult to implement across an entire application. What for example happens when you save? Can you never use EF directly? Is everything thread safe? You might want to simplify what you are trying to achieve.
In your example uow can be scoped to the HttpRequest. Many IoC containers (eg Structuremap) provide a simple way to configure this. You can then have a post-action filter (or even better an OWIN module) to attempt the commit (what happens if there are errors is yet another implementation difficulty to deal with). This will eliminate a lot of the property-assignment nonsense
I'm not sure what type of object is your StoreFrontLogic. It doesn't seem like a domain entity but it contains significant business logic. It could be something similar to a transaction script, but in that case the uow should be fully internal to it.
Is it a stateless service? In that case everything that method uses - orders for user included - should be passed in via a parameter.
If, on the other hand it's an entity then it shouldn't access the database at all, it should already have all orders for the user. Purposeful database denormalization can help quite a bit here.
At the very least pass uow as a parameter to OffersForUser rather than expecting for a property to be set.
I'm trying to build a new application using the Repository pattern for the first time and I'm a little confused about using a Repository. Suppose I have the following classes:
public class Ticket
{
}
public class User
{
public List<Ticket>AssignedTickets { get; set; }
}
public class Group
{
public List<User> GroupMembers { get;set; }
public List<Ticket> GroupAssignedTickets { get;set; }
}
I need a methods that can populate these collections by fetching data from the database.
I'm confused as to which associated Repository class I should put those methods in. Should I design my repositories so that everything returning type T goes in the repository for type T as such?
public class TicketRepository
{
public List<Ticket> GetTicketsForGroup(Group g) { }
public List<Ticket> GetTicketsForUser(User u) { }
}
public class UserRepository
{
public List<User> GetMembersForGroup(Group g) { }
}
The obvious drawback I see here is that I need to start instantiating a lot of repositories. What if my User also has assigned Widgets, Fidgets, and Lidgets? When I populate a User, I need to instantiate a WidgetRepository, a FidgetRepository, and a LidgetRepository all to populate a single user.
Alternatively, do I construct my repository so that everything requesting based on type T is lumped into the repository for type T as listed below?
public class GroupRepository
{
public List<Ticket> GetTickets(Group g) { }
public List<User> GetMembers(Group g) { }
}
public class UserRepository
{
public List<Ticket> GetTickets(User u) { }
}
The advantage I see here is that if I now need my user to have a collection of Widgets, Fidgets, and Lidgets, I just add the necessary methods to the UserRepository pattern and don't need to instantiate a bunch of different repository classes every time I want to create a user, but now I've scattered the concerns for a user across several different repositories.
I'm really not sure which way is right, if any. Any suggestions?
The repository pattern can help you to:
Put things that change for the same reason together
As well as
Separate things that change for different reasons
On the whole, I would expect a "User Repository" to be a repository for obtaining users. Ideally, it would be the only repository that you can use to obtain users, because if you change stuff, like user tables or the user domain model, you would only need to change the user repository. If you have methods on many repositories for obtaining a user, they would all need to change.
Limiting the impact of change is good, because change is inevitable.
As for instantiating many repositories, using a dependency injection tool such as Ninject or Unity to supply the repositories, or using a repository factory, can reduce new-ing up lots of repositories.
Finally, you can take a look at the concept of Domain Driven Design to find out more about the key purpose behind domain models and repositories (and also about aggregate roots, which are relevant to what you are doing).
Fascinating question with no right answer. This might be a better fit for programmers.stackexchange.com rather than stackoverflow.com. Here are my thoughts:
Don't worry about creating too many repositories. They are basically stateless objects so it isn't like you will use too much memory. And it shouldn't be a significant burden to the programmer, even in your example.
The real benefit of repositories is for mocking the repository for unit testing. Consider splitting them up based on what is simplest for the unit tests, to make the dependency injection simple and clear. I've seen cases where every query is a repository (they call those "queries" instead of repositories). And other cases where there is one repository for everything.
As it turns out, the first option was the more practical option in this case. There were a few reasons for this:
1) When making changes to a type and its associated repository (assume Ticket), it was far easier to modify the Ticket and TicketRepository in one place than to chase down every method in every repository that used a Ticket.
2) When I attempted to use interfaces to dictate the type of queues each repository could pull, I ran into issues where a single repository couldn't implement an generic interface using type T multiple times with the only differentiation in interface method implementation being the parameter type.
3) I access data from SharePoint and a database in my implementation, and created two abstract classes to provide data tools to the concrete repositories for either Sharepoint or SQL Server. Assume that in the example above Users come from Sharepoint while Tickets come from a database. Using my model I would not be able to use these abstract classes, as the group would have to inherit from both my Sharepoint abstract class and my SQL abstract class. C# does not support multiple inheritance of abstract classes. However, if I'm grouping all Ticket-related behaviours into a TicketRepository and all User-related behaviours into a UserRepository, each repository only needs access to one type of underlying data source (SQL or Sharepoint, respectively).
I have a program that deals with (virtual) money order transactions, and there are many classes that has logic requiring data to be saved to a database. My question is if I shall call my repositories directly from the various classes, or if I shall instead raise events, which a DatabaseManager class object can listen to, and hence all repositories will be called from this one class.
I don't have experience working with databases and repositories, so would appreciate some deeper insight and tips here. Like on what criteria you would chose different approach etc.
It's probably important to note that the database in this case is not used to retrieve data for performing program logic, except on program startup. So it's basically keeping all data in runtime objects, and just dumping to database for archiving.
I would pass an IRepository to my classes which they can then call to save data. For one thing it makes testing easier because you can easily inject a mock repository and on the other hand it makes it explicit that your classes have a dependency like that. You might want to search for the term Dependency Injection.
Simple example:
class Account
{
public Account(IRepository<Account> repository)
{
_Repository = repository;
}
public void ChangeOwner(Owner newOwner)
{
// change ownership
_Repository.Save(this);
}
}