Pages

Thursday 25 April 2013

Types of Entities in Entity Framework: EntityObject, POCO, POCO Proxy and Self-Tracking Entities


Types of Entities in Entity Framework: EntityObject, POCO, POCO Proxy and Self-Tracking Entities

There are four types of Entities in Entity Framework: 

1) EntityObject 
2) POCO (Plain Old CLR Object)
3) POCO Proxy 
4) Self-Tracking Entities

EntityObject

By default, the ADO.NET Entity Data Model tools generate EntityObject derived entities. When you work with EntityObject derived types, the object context manages the relationships between your objects, tracks changes as they occur, and supports lazy loading in the most efficient manner. However, the EntityObject derived types have strong dependency on the Entity Framework.

POCO (Plain Old CLR Object)

POCO class is the class which doesn’t depend on any framework unlike EntityObject specific base class. It is like any other normal .net class that is why it is called “Plain Old CLR Objects”. The Entity Framework enables you to use existing .net classes together with the data model without making any modifications to the existing .net classes. These POCO entities (also known as persistence-ignorant objects) support most of the same LINQ queries as EntityObject derived entities.

POCO Proxy

POCO Proxy is a runtime proxy class of POCO entity. POCO entity becomes POCO Proxy entity if it meets certain requirements to enable lazy loading proxy and instant change tracking. It adds some methods at runtime to your POCO class which does instant change tracking and lazy loading stuff. 

POCO entity should meet the following requirement to become POCO proxy:

1. A custom data class must be declared with public access.
2. A custom data class must not be sealed (Not Inheritable in Visual Basic)
3. A custom data class must not be abstract (Must Inherit in Visual Basic).
4. A custom data class must have a public or protected constructor that does not have parameters.
5. The class cannot implement the IEntityWithChangeTracker or IEntityWithRelationships interfaces because the proxy classes implement these interfaces.
6. The ProxyCreationEnabled option must be set to true.
7. Each navigation property must be declared as public, virtual

Difference between POCO and POCO Proxy Classes

Main difference between POCO and POCO Proxy lies in save changes mechanism. POCO entities use Snapshot mechanism. So before saving changes POCO entities, you must use DetechChanges method of ObjectContext to synchronize data with ObjectStateManager. Whereas POCO Proxy entities use instant change tracking. It synchronize with ObjectStateManager as soon as something changes in the POCO Proxy entity. So it’s more efficient than snapshot mechanism and you don’t have to use DetechChanges method of ObjectContext. Thus POCO Proxy has a same functionality as EntityObject.

Self-Tracking Entities

The EntityObject derived entities, POCO, and POCO proxy entities works well in application where the entity objects can be attached to the object context that handles change tracking on single tier. However, in n-tier application, you have to transfer entities to a tier where the object context is not available e.g. Business tier or presentation tier. So how to track changes and report those changes back to the object context? Answer is “self-tracking entities”. Self-tracking entities as its name suggests, can record changes to scalar, complex, and navigation properties on its own. Self-tracking entities do not depend on the Entity Framework so it can be transferred to other tier.  

Self-Tracking entities have additional change tracking functions and it implements IObjectWithChangeTracker and INotifyPropertyChanged interface. It also mark it as DataContract to be used in WCF services.

No comments:

Post a Comment

About the Author

I have more than 10 years of experience in IT industry. Linkedin Profile

I am currently messing up with neural networks in deep learning. I am learning Python, TensorFlow and Keras.

Author: I am an author of a book on deep learning.

Quiz: I run an online quiz on machine learning and deep learning.