I implemented Windows Identity into my application and using EF Migrations the appropriate [AspNet*] tables were added to my database. As typical in this scenario, I created an ApplicationUser class that inherits from IdentityUser. My 'ApplicationUser' class does not have any additional properties. And at that point in time I did not touch any code or classes related to Roles.
Now I am looking to add roles so I followed the same pattern as before. I created an 'ApplicationRole' class that inherits from 'IdentityRole' but did not add any addition properties.
However, and this is where my question comes in, EF migrations is trying to create a [Discriminator] column for the AspNetRoles table. Yet it never created one for the AspNetUsers table in the previous migration.
I understand the purpose of the Discriminator column so I am not questioning why it is there. I just do not understand why it is adding it for one table and not the other when both base classes were inherited by the "application" equivalent classes.
The only plausible reason I can think of is because the original IdentityRole class was used in the earlier migration and now I am trying to add a new migration that uses the base class. If so, is there a way to get EF to ignore this? There are no plans to ever create multiple classes that inherit from the base nor will I be using the base directly so the Discriminator will always contain the same value.
Related
I want to use the EF Core HasData method to seed a database with reference data. Two of the models I want to seed follow the Table per Hierarchy (TPH) pattern in the sense that there is a none abstract base type and another none abstract derived type. These models in the database exist in a table named after the base type with a discriminator column. I would like to add data seeding for this table. However I’m struggling to find any guidelines of how to do so in the EF Core 6 documentation.
I have a few related questions:
Do I seed all the data using the HasData method on an EntityTypeBuilder<BaseType> or do I need to split the seeding into one HasData on that class and another on EntityTypeBuilder<DerivedType>?
I understand using TPH will add a shadow discriminator property and that, potentially, I’ll have to add that in the data seeding. Does that mean I have to use anonymous types to specify that property (doesn’t seem very elegant) and if so, can I get the autogenerated discriminator name using a method (typing it manually sounds like a risk as what if EF Core changes the discriminator name convention?)?
Should I be avoiding using TPH on reference tables altogether? Is there something else I should do instead?
After going through my options with trial and error, I have come to the following solution. It may not be perfect, but it deals with all of my concerns.
Do I seed all the data using the HasData method on an EntityTypeBuilder<BaseType> or do I need to split the seeding into one HasData on that class and another on EntityTypeBuilder<DerivedType>?
You have to do it on both classes, attempting otherwise throws on creating the migration:
The seed entity for entity type 'BaseType' cannot be added because the value provided is of a derived type 'DerivedType'. Add the derived seed entities to the corresponding entity type.
I understand using TPH will add a shadow discriminator property and that, potentially, I’ll have to add that in the data seeding. Does that mean I have to use anonymous types to specify that property (doesn’t seem very elegant) and if so, can I get the autogenerated discriminator name using a method (typing it manually sounds like a risk as what if EF Core changes the discriminator name convention?)?
I have added a property to the BaseType class and defined it as the discriminator, so now I can specify the discriminator values directly without using anonymous type and I am in control of the discriminator values. Gert Arnold points out in this other answer why this might not be appropriate. However I have used the following method described in this EF Core Github issue to hopefully mitigate the concerns he raised.
Should I be avoiding using TPH on reference tables altogether? Is there something else I should do instead?
This is still unanswered. But I am feeling a lot more confident it is a good approach now.
We've been working for along time using EF Designer from database, so switching to another approach is not an option for us.
Due to a legacy version of our app i need to create relationship between tables on EF level for the current version, so those relationships don't exist in the schema.
I understand that i can do do that using Code First from database approach, but the problem how do i map the fields and make these entities part of the main datacontext which we already extended its functionalities using another partial class.. the main problem now is just mapping the new entities fields
i can't create OnCreatingModel in a new datacontext class since it does exist in the auto generated datacontext class.
to sum up.. is there anyway to use both approaches on the same DAL? taking in mind modifying auto-generated classes is absolutely not an option.
Thanks!
I'm using Entity Framework for the first time. I'm using Code First development with an existing database. Everything is working great but I do not like using the table name as the class name. The database tables do not making for friendly class names in C#. Is it common place that people rename the classes that are auto-generated by EF? Will this cause problems down the line in some capacity that I am not anticipating? FWIW, I'm not worried about future schema changes to the tables.
I Google'd a few times but no one seems to have asked this before. Maybe I should take that as a sign....
You can use either data annotation attributes:
[Table("tblFoo")]
public class YourEntity
{
}
Or fluent api
modelBuilder.Entity<YourEntity>()
.ToTable("tblFoo");
to provide table name which entity maps to.
Im currently working in a team that uses EF as the ORM of choice.
We have a common project that contains many EDMX files.
The reason for this is to keep the EDMX files small and manageable while also allowing them to focus on a conceptual set of tables on the database.
Eg
Orders.edmx
Users.edmx
Trades.edmx
These all point to a different set of tables on the same db.
I now need to add the user table to the Trade.edmx file. Since the user table is already in the user.edmx file, this creates the same User type twice under a different namespace which means I would need 2 UserRepository objects.
Common.data.trade.User
Common.data.users.User
Is there a way of avoiding 2 repository objects for the same table?
Any suggestions would be greatly appreciated
If you are using POCO generator you can update template for Trades.edmx to not generate new User class and its context template to use User class from Users namespace. EF matches POCO classes with entities in designer only by the class name (namespace is omitted) so it will work.
The disadvantage is that you have User entity in two mapping files and you must update it in both files or your application throw exception at runtime.
The reason for this problem is your architecture - at the beginning you wanted separated models but know you want to combine entities from different models. Those are contradicting requirements. Either use separated model where Trade knows only userID without any navigation property (like if it is defined in another database) or move all entities to single EDMX to support your new requirements.
Our organization is looking to standardize on Entity Framework once v4 comes out. As a result, I am looking at what it would take to migrate our application that uses NHibernate for persistence to EF4 using POCO support. In a couple of places we use single table inheritance (also known as Table Per Hierarchy). I have been unable to get it to work using the following.
Payment (base class [should be abstract but having trouble there as well])
CreditCardPayment (concrete implementation)
ACHPayment (concrete implementation)
CheckPayment (concrete implementation)
Right now, I am mapping them with only the base class properties. All of these classes are in the same namespace. They have a discrimimator that is called PaymentTypeId in the database, so the Payment mapping has a condition of "When PaymentTypeId = 0". Each of the subclasses have the same condition with different values (i.e. CreditCardPayment = 1, etc.).
When I try to load each a list of all payments using DataContext.Payments.ToList() (DataContext inherits from ObjectContext) I am getting the following exception:
"Object mapping could not be found for Type with identity 'DataLayer.DataModel.CreditCardPayment'."
I can't figure out what this means, as the POCO CreditCardPayment class lives in the same namespace as the POCO Payment class does (in fact in the same file).
What am I missing?
This is complaining not about database mapping, but model to CLR mapping.
The EF can't for some reason find your CreditCardPayment class.
Now one possible reason is that you haven't loaded the metadata for it yet.
For example if you have this:
Assembly 1:
- Payment
Assembly 2 references Assembly 1:
- CreditCardPayment extends Payment
Then when you query the EF doesn't know where CreditCardPayment lives.
The way to get around this is with LoadAssembly i.e:
using (DataContext ctx = new DataContext())
{
ctx.MetadataWorkspace.LoadFromAssembly(typeof(CreditCardPayment).Assembly);
// now do your query.
}
You need to tell to LoadFromAssembly every assembly that isn't directly reference by your DataContext class.
Note: typeof(Payment).Assembly is directly referenced because of the IQueryable<Payment> Payments property.
Hope this helps
Alex
Microsoft.
What I didn't represent before (I didn't think it relevant, but it was). Was that CreditCardPayment inherited from an intermediary class named "CreditPayment" and ACHPayment inherited from CashPayment. CreditCardPayment and CashPayment live in the same namespace and file, but were not represented in the EF mapping. Once I added those within the mapping file, everything worked ok.
So, even thought EF does not ever map to one of those types directly, it seems to need to know about them, because it changes the basetype of the CreditCardPayment classes et al. Thank you for your help on this.