In a code-first implementation of EF, used along with a Repository pattern, should the repository return business objects from the domain model, or simple entities from the data model?
From what I can tell, the point of Repository is to return business objects, not entities, so that you can do work with them. But most of the code examples I find are returning data models instead, which seems like a bad idea to me because what if the data source changes?
If you are using CodeFirst development you can easily use your business object as the data model as well. You can write Ef mappings in a separate DLL to remove the dependency of EF to business model. If you want to change the data source to another one instead of EF you can keep the same Business(domain) classes for that as well.
Related
I have a domain model architecture in which my domain/business objects were created based on the problem domain and independent of any knowledge of the physical data model or persistence structures. So far I'm on track because it's perfectly acceptable and often the case that there is an impedance mismatch between the domain model and the data model. A DBA created the database for getting the data they required, but it does not encapsulate the applications entire domain model or design.
The result - I have my own set of domain model objects. However all of the fields that need to be persisted do exist somewhere or another within my domain model, but not necessarily in the shape that my auto generated .edmx POCO entities have them. So I have all the data, it's just not in the perfect shape exactly like the tables in which auto generated POCO entities are created from.
I have seen a few posts on this topic like converting POCO entity to business entity and Entity Framework 4 with Existing Domain Model that make statements like the following:
"Create the entities in your entity data model with the same names as
your domain classes. The entity properties should also have the same
names and types as in the domain classes"
What!? No way, why should I have to make my domain model be reshaped to POCOs that are modeled exactly after the data model / table structure in the database? For example - in my case of having 5 given properties, 2 might be in class 'A' and 3 in class 'B', whereas a auto generated POCO class has all 5 in its own class 'A'.
This is the entire point, I want separation of my object model and data model but yet still use an ORM like EF 5.0 to map in between them. I do not want to have to create and shape classes and properties named as such in the data model.
Right now my .edmx in EF 5.0 is generating the POCO classes for me, but my question is how to dissolve these and rewire everything to my domain objects that contain all this data but just in a different shape?
By the way any solution proposed using a Code First approach is not an option so please do not offer this. I need some guidance or a tutorial (best) using EF5 (if possible because EF4 examples are always inheriting POCOs from ObjectContext) with wiring up my own business objects to the .edmx.
Any help or guidance is appreciated, thanks!
This sounds like exactly the use case of Entity Framework. I am making a few assumptions here. First that when you make this statement:
"I have a domain model architecture in which my domain/business objects were created based on the problem domain and independent of any knowledge of the physical data model or persistence structures."
That you mean this domain was created in the EF designer? But then you say:
"However all of the fields that need to be persisted do exist somewhere or another within my domain model, but not necessarily in the shape that my auto generated .edmx POCO entities have them."
This sounds to me like my first assumption is incorrect.
Next, you dismiss code first? If your domain model/business objects are code based and you want to persist them to a relational database, that is exactly the use case for code first. You have the code, now you need to create your DbContext and map it to your physical model.
However you dismiss that... so some thoughts:
If you have a domain model of code based business objects and you have an EDMX that is used for other things I think you would want to create a repository layer that uses something like auto mapper or manual projections to query your Entities and return your business objects.
If you have a domain model of code based business objects and you have an EDMX that is not used for other things other than persisting your business objects I would say that you need to express your domain in an EDMX and then map it onto your existing database. This is really the use case for an ORM. Having two domain models and mapping from one model to the other where one model matches your domain and one matches your database is adding an extra un-needed layer of plumbing.
The latter approach above is what is called "Model First" in EF parlance. There are several articles written about it although the bulk of them just generate the db from the model. You would not do that step, rather you would map your entities onto your existing database.
The basic steps for this are to "update from the database" not selecting any of the db objects (or entities would be created). Or, you can take your exiting .edmx in the designer (which is sounds like you have) and modify the entities to match your business domain. Or just delete all the entities in your EDMX model, create your entities as you want them, and then map them all.
Here is a jing I made where I use the EF Designer to bring in the model store (the only way to do this is to allow it to generate entities) and then delete the entities allowing the Store information to stay by clicking NO when it asks if you want to delete the table info.
http://screencast.com/t/8eiPg2kcp
I didn't add the POCO generator to this, but if I did it would generate the Entities in the designer as POCO classes.
The statement quoted above is not suggesting that you rewrite your domain objects to match your pocos, it is suggesting that you update the edmx to match your domain model.
In your example you could create an entity in your edmx that maps all 5 properties from both tables and EF will manage the mapping to and from the single generated Poco onto your tables.
Of course this means that you then have duplicate domain objects and pocos, meaning you would either have to manually convert your domain objects to pocos when interacting with EF,
or you could define your domain data objects as interfaces and provide partial implementations of the pocos that essentially identify each poco as being a concrete implementation of a domain object.
There are probably several other ways to skin this particular cat, but I don't think that you can provide predefined c# objects for use in an edmx.
One approach might be to select into a ViewModel (suited to your particular business logic) and automatically map some data from the context into it like so https://stackoverflow.com/a/8588843/201648. This uses AutoMapper to map entity properties from an EF context into a ViewModel. This won't do everything for you, but it might make life a bit easier. If you're unhappy with the way this occurs automatically, you can configure AutoMapper to do things a bit differently (see Projection) - https://github.com/AutoMapper/AutoMapper/wiki/Projection
You might know this already, but its also possible to automatically generate POCOs from your EDMX using t4 - http://visualstudiogallery.msdn.microsoft.com/72a60b14-1581-4b9b-89f2-846072eff19d. If you define the templates to generate partial classes, you can then have another partial class with your custom properties for that POCO. That way you can have most properties automatically populated, but have other custom properties which you populate with custom rules from your context/repository. This takes a lot of the monotony out of generating these, and you can then focus on reshaping the data using the above technique.
If you're seriously unhappy with both, you could always map a stored procedure to get the exact field names that you want automatically without needing to stuff around. This will of course affect how you work with the data, but I have done it before for optimisation purposes/where a procedure already existed that did exactly what I wanted. See http://msdn.microsoft.com/en-us/data/gg699321.aspx
I'm new to the MVC way of developing applications and for the most part am enjoying. One thing I'm a bit confused about is the use of the Entity Framework. The EF usually (at least in my experience) defines multiple tables and relationships through the .edmx table. A couple of questions:
Why would I define a separate class file for a specific table if EF is building all of the classes that I need in the background?
From some of the validation approaches that I've seen, they want to define validation logic in the class related to a model for a table. If I'm using EF, will I have a .cs file describing the model and a .edmx describing that same table (in addition to its associated tables)?
If yes, how do you connect the .cs file to the .edmx definition so that CRUD flows easily from the EF?
Sorry if these seem like easy questions but I'm just trying to get my head wrapped around these fundamental concepts. Too many examples out there use only a single table where in my business, I NEVER write an application that uses a single table. There are always multiple tables in relation to each other with foreign keys. Thanks for your prompt responses.
For a tutorial that shows the use of partial classes -- in a Web Forms application but for MVC the same technique would be used -- see Adding Metadata to the Data Model in this tutorial:
http://www.asp.net/web-forms/tutorials/getting-started-with-ef/the-entity-framework-and-aspnet-getting-started-part-8
From your comment "The EF usually (at least in my experience) defines multiple tables and relationships through the .edmx table." it sounds like you are familiar only with Database First and Model First -- for an introduction to Code First and an explanation of the differences, followed by a series of tutorials with an MVC example using Code First, see this tutorial:
http://www.asp.net/mvc/tutorials/getting-started-with-ef-using-mvc/creating-an-entity-framework-data-model-for-an-asp-net-mvc-application
Good questions, Darryl. Here are my responses to your bullet points:
Defining separate model classes that match the data models that EF creates is generally a good idea for the simple sake of separating your data access "stuff" from your business model objects that will get used throughout your app. Some people don't like this approach because it creates some amount of overhead when it comes to mapping your entities to POCOs but, if you use a tool such as AutoMapper, the overhead is minimal. The benefit lies in you creating a layer of separation between you and your (likely) evolving data model.
You could define validation logic in a buddy class (just a partial class that sits along-side your entity) but that would mean that you would be using that entity across your app and some would debate that that isn't the best idea. The alternative method, as mentioned above, is to create your own POCOs to mirror the entities that EF creates and place your validation attributes on the POCOs.
I mentioned this in the previous item but the way to do this would be to define buddy classes. Give EF buddy classes a Google and you should find plenty of examples on how to do that.
Just to add to all of this, if you choose to create POCO classes that mirror your EF entities, tools like AutoMapper can handle fairly complex relationships when it comes to mapping classes. So, if you have foreign key relationships in your data model, AutoMapper can understand that and map your POCO classes accordingly (i.e.: You have an entity that has a 1-to-many relationship and a POCO with a list of objects to mirror that relationship.)
I hope some of that helps...
I know that entity framework has a database first approach. Now the question is whether it can generate the DAL (data access layer) code (not the models) for me.
When using a Object Relational Mapper (ORM), you don't typically have CRUD code in the traditional sense. Rather, it abstracts those operations into more object oriented operations.
For example, you don't "insert", you add the model class to the table, then save changes. The ORM automatically generates the SQL needed to make the Object model match the data model.
So my point is, your question displays a basic lack of understanding of how ORM's work and how they relate to data models. You should probably do a little reading.
I'm not sure what you mean specifically by "DAL code", as that's a rather ambiguous term. I would consider your Entity types part of the DAL.
When you use a model-first or database-first approach, the Entity Framework tools can auto-generate a context class from your model .edmx, which will inherit from ObjectContext. It's easy to customize the generated context class with T4 templates by finding one online that already generates from a .edmx, and modifying to your liking.
Code-first development uses the DbContext, which is not typically auto-generated. Please see this post on Scott Gu's blog for more details on this.
We are in a process of designing an application with approx 100 tables and complicated business logic. Windows Forms will be used on the client side and WCF services with MSSQL on the server.
Custom DTOs are used for client-server communication, business entities are not distributed.
Which variant of Entity Framework to use (and why):
EF 4.0 EntityObjects
EF 4.0 POCO
EF 4.1 DbContext
Something else
Database-first approach is a requirement.
Also, is it worth implementing a Repository pattern? It seems a bit redundant, as there is one level of abstraction in the mapping itself and another one in the use of DTOs. I'm currently leaned towards using auto-generated extendable repositories for each entity returning IQueryable, just to have a place to put common queries, but still allowing querying entity model directly from the Service Layer.
Which variant to use? Basically once you have custom DTO the only question is do you want to have control over entities code (their base class) and make them independent on EF? Do you want to use code first? If the answers to all questions are no then you can use EntityObjects. If you want to have entities persistence ignorant or use custom base class you should go to POCO. If you want to use code first or new DbContext API you will need EF 4.1. Some related topics:
EF 4.1 Code-first vs Model/Database-first
EF POCO code only VS EF POCO with Entity Data Model (this was related to CTP)
ADO.NET DbContext Generator vs. ADO.NET POCO Entity Generator
EF Model First or Code First Approach?
There are more things to consider when designing service layer. You should be aware of complications you will have to deal with when using EF in WCF. Your service will provide data to WinForms application and it will work with them in "detached mode". Once user will do all changes he wants to do he will post data back to the service. But here comes the problem - you must tell EF what has changed. If you for example allow user to change order with all its order items (change quantity in items, add new items, delete some items) you must say EF exactly what has changed, what was added and what was deleted. That is easy when you work with single entity but once you allow user to change object graph (especially many-to-many relations) then it is quite tough. The most common solution is loading the whole graph and merge the state from incoming DTOs to loaded and attached graph. Other solution is using Self tracking entities instead of EntityObjects/POCOs + DTOs.
When discussing repositories I would refer you to this answer which refers many other answers discussing repositories, their possible redundancy and possible mistakes when using them just to make your code testable. Generally each layer should be added only if there is real need for the layer - due to better separation of concerns.
The main advantage of POCOs is that those classes can be your DTOs, so if you've already got custom DTOs that you're using, POCO seems a bit redundant. However, there are some other advantages which may or may not have value to you, since you didn't mention unit testing as a requirement. If you plan to write unit tests, then POCO is still the way to go. You probably won't notice much difference between 4.0 POCO and 4.1 since you won't be using the code-first feature (disclaimer: I've only used 4.0 POCO, so I'm not intimately familiar with any minor differences between the two, but they seem to be more or less the same--basically I was already using POCO in 4.0 and haven't seen anything that's made me want to update everything to use 4.1).
Also, depending on whether you plan to unit-test this layer, there's still value in implementing the repository/unit of work patterns when using Entity Framework. It serves to abstract away the data access logic (the context), not the entities themselves, and allows you to do things like mocking your context in unit tests. What I do is copy the T4 template for my context and use it to create the interface, then edit the T4 template for the context and have it implement that interface and use IObjectSet<T> instead of ObjectSet<T>. So instead of:
public class MyEntitiesContext
{
public ObjectSet<MyClass> MyEntities
...
}
I end up with:
public interface IMyEntitiesContext
{
public IObjectSet<MyClass> MyEntities;
}
and
public class MyEntitiesContext : IMyEntitiesContext
{
public IObjectSet<MyClass> MyEntities
...
}
So I guess it really comes down to whether or not you plan to write unit tests for this layer. If you won't be doing anything that would require mocking out your context for testing, then the easiest thing to use would probably be 4.0 EntityObjects, since you aren't planning to pass your entities between layers and it would require the least effort to implement. If you plan to use mocking, then you'll probably want to use POCO and implement repository/unit of work.
I have not hardly touched EF4, but I've used Linq to sql quite a lot. I would like to start into one of the EF templates but I have no idea what situations make sense for each or what their intent was.
I have the following possibilities:
Data templates
ADO.NET Entity Data Model
Service-based Database (is this even related to EF?
Code templates (I am familiar with T4)
ADO.NET EntityObject Generator
ADO.NET Self-Tracking Entity Generator
Online Templates
ADO.NET C# POCO Entity Generator
I have no idea what situations make
sense for each or what their intent
was
Not meaning to sound rude, but did you have a look on MSDN/ASP.NET to find out? There is plenty of information around. And there is a lot to each of those templates, more than i can go into here. There is a MSDN page for each of these.
That being said, i'll give you a quick summary, so people who stumble here have some info.
ADO.NET Entity Data Model
This is the file you create to use Entity Framework as your ORM, and it is mandatory for using EF. You need this before you use any of the others. You can create your EDM with a number of different approaches, including database-first (generate from DB), code-first, model-first, etc.
Service-based Database
I have never heard of this term, and given i've been working with EF a lot lately (and reading), i doubt this will be related to EF.
ADO.NET EntityObject Generator
Generates classes for entities which inherit from the EntityObject class. Identical to the default EF code generator, except instead of putting output code into the Model.edmx.designer.cs (default) file, the code gets put into seperate files. I personally don't see any benefit in this template.
ADO.NET Self-Tracking Entity Generator
Generates classes for entities when you want to develop N-Tier applications (ie if you wanted to allow a WCF/Silverlight app to work with your model). Entities are setup to be 'trackable' by the EF Graph, in order to handle persistence operations from various applications.
ADO.NET C# POCO Entity Generator
My favourite. :) Generates classes for entities which inherit from nothing. They have no idea that they are being used for persistence. Use this for applications when you want persistence-ignorance, testability and loose-coupling of your domain/persistence layers.