I am working on creating a Technical Design Document for my new project that is being developed using Entity Framework 4.1 Code First.
I want my DAL to be to be loosely coupled, easily testable and should be able to dynamically inject it using IoC Container...
Thinking of using Repository Pattern. Are there any good resources for reference?
--
Preetham Reddy
See these tutorials:
http://www.asp.net/entity-framework/tutorials/implementing-the-repository-and-unit-of-work-patterns-in-an-asp-net-mvc-application
http://www.asp.net/entity-framework/tutorials/using-the-entity-framework-and-the-objectdatasource-control-part-2-adding-a-business-logic-layer-and-unit-tests
This is a pretty good post.
http://blogs.msdn.com/b/adonet/archive/2009/06/16/using-repository-and-unit-of-work-patterns-with-entity-framework-4-0.aspx
Related
I am new to entity framework and MVC, and trying to understand what constitutes a good design approach for a new application.
There are several ways of using Entity Framework. However, for my project, the best looking option is DB First. I've played around with an EDMX file, and I have got as far as using the DbContext code generator to create my wrapper classes.
I plan on using the repository and unit-of-work patterns, and using ninject for DI.
However, it does not seem "proper", from a SoC point of view, that whilst my respository will hide the implementation of the data store (EF) from my code, the model classes themselves are very much EF flavoured.
It seems that using EDMX-based approaches to EF blur the separation of concerns. Only POCO support seems to allow a true separation, but POCO has some other limitations that I don't like.
Am I missing something, or does using EDMX have this drawback?
Are people using an auto mapper to convert between the entity model and another, clean, SoCced model?
thanks
Tian
I don't have a strong opinion on the Separation Of Concerns question, but I have used both the standard ADO.Net version of EF and POCO and it is not difficult at all to customise the output of the the T4 code generation script for POCO to address any concerns you have about the structure of the objects created. That sounds like it would probably be a good starting point for what you are looking to do.
Once you know you are looking for T4 templates there are quite a few tutorials and a lot of helpful SO questions that can give you an idea of what you need to do.
I'm developping a line of business application using WPF as a presentation layer (of course with MVVM).
I'm using ADO.Net Entity Framework to map the DataBase.
I don't want to use entities directly in code (in the business layer). I want to separate my project into 3 layers:
Presentation layer
Business Layer
Data Access Layer
According to this post I want to implement a full encapsulation of the Entity Framework to provide a separation of concerns and to not be dependant on EF as ORM in the future.
Can you help me by giving me some exemples to encapsulate the EF and how to implement this in code.
Regarding this
I want to implement a Full encapsulation of the Entity Framework. to
provide a separation of concerns and to not be dependant on EF in the
future as ORM
Normally, you will create yourself a lot of problems if you go that route. If you choose EF, you really should make full use of the features, not hiding that behind another abstraction.
EF itself is already an abstraction layer over DB, there is no need to create another abstraction on top of that.
I would take a look at this post wich implements UnitOfWork and Repository patterns to implement what, I understand, you want to achieve.
http://blogs.msdn.com/b/adonet/archive/2009/06/16/using-repository-and-unit-of-work-patterns-with-entity-framework-4-0.aspx
There is one way of doing it, using POCO. Entity Framework 4.0 comes with the support of POCO (Plain CLR Objects). But POCO has its own complexities, when u have to deal with Relationship and associations. You can refer to the blog by Julie Lerman (a nice article)
http://thedatafarm.com/blog/data-access/agile-entity-framework-4-repository-part-1-model-and-poco-classes/
i developed an sample application using Entity Framework 4.0. but now i'm gonne developed a real app for a company. where i would like to implement the DAL with Entity framework with generic repository , unit of work and DI container.
please any one suggest me a real example for the framework..
Thanks
Rusho
Generic repository is nonsense. If you want to use design pattern called Repository you should think about specific repository and aggregate roots.
Generic repository is just a wrapper around ObjectSet / DbSet providing no added value - only additional layer which must be maintained and which makes interaction with EF harder. Also adding repository without clarifying why you want to do that and what it should solve for you is wrong approach - design pattern is a blue print for solving a problem. Not something you should use just because it exists and everybody talks about it.
You can also check these answers where I discuss generic repository and its implications:
Generic Repository With EF 4.1 what is the point
The repository itself is not usually tested?
Reference To This Question
Where one should put the DQL queries? In a service class, or in a controller or maybe in a repository class?
Found a nice article about this, which answers my question. I think it's best to put them in the service class.
How to integrate Doctrine 2 with your Zend Framework application
This highly depends on what your DQLs are doing:
If you have a Query which is doing work on only one entity type I suggest to create your own Repository class for this entity. The repository class already provides you with the methods for find and findAll, so it would fit there good.
Doctrine gives you orm:generate-repositories as CLI tool. Ralph Schindler took this approach as you can see in his example repository.
If you have a Query which affects multiply types of entities, the Service Layer should be the best place to put it.
I currently use the Entity Framework designer to generate my persistance objects and I also user POCO view models for the ASp.NET MVC view.
I've read and listened to a lot of people talking about the good support for POCO's in EF4 as well as POCO's in general but I can't seem to work out what advantage, if any I'll get from using them.
In our application, we WILL be using SQL Server so it's not like we need so separate out for different databases.
Why would I want to use POCO's as opposed to the designer generated classes?
POCO offers better extensibility/reuse of your Domain Model as you're not tied to any specific ORM framework.
Answered here :What are the 'big' advantages to have Poco with ORM?
Easier to unit test
I find when you have many entities (100+) using the designer is painful and POCO objects are easy to create and maintain.
With POCO's you can "Code First"
http://weblogs.asp.net/scottgu/archive/2010/08/03/using-ef-code-first-with-an-existing-database.aspx
+1 for with POCOS you can code first. In fact you can create and test your entire application before you even write your first line of data access code. This is how it should be, it's a perfect application of the SRP. Your domain object should not know or care how they are being persisted.