How many domain models should I create

Thread is closed for posting
2 posts, 0 answers
  1. Phil
    Phil avatar
    1 posts
    Member since:
    Dec 2013

    Posted 07 Mar 2014 Link to this post

    I am looking for some help in creating my domain model(s). I am both new to Telerik Data Access and database modeling. I have a legacy database with 450 tables and views. We are using fluent mapping to create the domain model. Is it better to only have one domain model? I think breaking the tables up into logical collections of tables makes more sense. Is there any reasons why this is not a good approach? One that I can see is that the same table may exist in multiple models such as an item table. We are developing an enterprise solution and this will be deployed on both Oracle and Sql Sever.
  2. Boyan
    Boyan avatar
    100 posts

    Posted 12 Mar 2014 Link to this post

    Hello Phil,

    In order to support different database servers you will have to use our code-only mapping as you are already doing. This documentation article provides more information on how to implement a switching mechanism for the database servers you need to support and also lists some points you will need to consider when building your fluent mapping model. If you are not planing to switch between database server during runtime you could just change the connection in your configuration file.

    On the matter of the multiple models, our general recommendation in such cases would be to use a single model. Although a valid approach, using multiple models could lead to higher memory usage and possibly synchronization issues is some situations. For those reason the multiple models approach is not suitable for logical separation of the database. If you have any additional consideration that could benefit from multiple models, please let us know so we could advice you better. 
    As an alternative, in order to separate your code logically, I would recommend you to group the fluent mapping code in separate files as dictated by you database schema. Each file will contain the mapping for specific tables. Then all those mappings could be used to create a new FluentMetadataSource that combines them. The resulting FluentMatadataSource could be used by the OpenAccessContext class. This way the database mapping will still be separated into logical categories in your code  but you will avoid any overhead that would not be required in your scenario.

    Should you have any further question or need any further assistance, please let us know.

    OpenAccess ORM is now Telerik Data Access. For more information on the new names, please, check out the Telerik Product Map.
Back to Top