I am using MVC3 and Entity Framework. I have a class called User with 20 different properties. I have already created a database and filled it with some data. I want to break out the Addresses property and make it it's own class.
namespace NameSpace.Domain.Entities
{
public class User
{
public int UserId { get; set; }
...
...
public string AddressOne { get; set; }
public string AddressTwo { get; set; }
}
}
I want to break out both Addresses like so
namespace NameSpace.Domain.Entities
{
public class User
{
public int UserId { get; set; }
...
...
public Addresses Addresses { get; set; }
}
public class Addresses
{
public string AddressOne { get; set; }
public string AddressTwo { get; set; }
}
}
HERE'S MY QUESTION:
Since I already have the data table filled with data, how can I update this in the Server Explorer?
Thanks ( if you need more info please let me know )
If you are using EF code first 4.3 you can use the concept of migrations to achive what you want.
You will need to do a code based manual migration since you change is a bit to advanced for the framework to figure it out itselfe.
Further reading: http://blogs.msdn.com/b/adonet/archive/2012/02/09/ef-4-3-code-based-migrations-walkthrough.aspx
Related
I've a situation where entity framework core (2.0) is performing additional work to a parent table when I update a row in a child table. I've pin-pointed the cause to a value not being set in the unflattened object tree produced by AutoMapper (I'm not saying it is an error in AutoMapper; it's probably more to do with my code).
I'm using ASP.NET Core 2.0, C#, EF Core 2.0 and AutoMapper for the API development side. The database already exists and the EF classes scaffolded from it.
To keep it short, the child table is Note and the parent table is NoteType. The EF classes (extraneous columns removed) are as follows :
//Entity classes
public partial class Note
{
public int NoteBookId { get; set; }
public short NoteSeqNo { get; set; }
public short NoteTypeId { get; set; }
public string NoteText { get; set; }
public NoteBook NoteBook { get; set; }
public NoteType NoteType { get; set; }
}
public partial class NoteType
{
public NoteType() { Note = new HashSet<Note>(); }
public short NoteTypeId { get; set; }
public string NoteTypeDesc { get; set; }
public ICollection<Note> Note { get; set; }
}
//DTO class
public class NoteDto
{
public int NoteBookId { get; set; }
public short NoteSeqNo { get; set; }
public short NoteTypeId { get; set; }
public string NoteTypeNoteTypeDesc { get; set; }
public string NoteText { get; set; }
}
public class NoteTypeDto
{
public short NoteTypeId { get; set; }
public string NoteTypeDesc { get; set; }
}
(NoteBookId + NoteSeqNo) is Note's primary key.
NoteTypeId is the NoteType's primary key.
Configuration
This is the AutoMapper configuration:
// config in startup.cs
config.CreateMap<Note,NoteDto>().ReverseMap();
config.CreateMap<NoteType,NoteTypeDto>().ReverseMap();
Read the data
As a result of data retrieval I get the expected result and the parent note type description is populated.
// EF get note in repository
return await _dbcontext.Note
.Where(n => n.NoteId == noteId && n.NoteSeqNo == noteSeqNo)
.Include(n => n.NoteType)
.FirstOrDefaultAsync();
// Get note function in API controller
Note note = await _repository.GetNoteAsync(id, seq);
NoteDto noteDto = Mapper.Map<NoteDto>(note);
Example JSON result:
{
"noteBookId": 29,
"noteSeqNo": 19,
"noteTypeId": 18,
"noteTypenoteTypeDesc": "ABCDE",
"noteText": "My notes here."
}
Update the data
When the process is reversed during an update, the API controller maps the dto to the entity
Mapper.Map<Note>(noteDto)
Then when it is passed to EF by the repository code, EF tries to add a NoteType row with id 0. The unflattened object tree looks like this:
Note
NoteBookId = 29
NoteSeqNo = 19
NoteTypeId = 18
NoteTypeNoteTypeDesc = "ABCDE"
NoteText = "My notes updated."
NoteType.NoteTypeDesc = "ABCDE"
NoteType.NoteTypeId = 0
The parent id column (NoteType.NoteTypeId) value is 0 and is not assigned the value of 18 which is what I expected.
(During debugging I manually set NoteType.NoteTypeId to 18 to ensure EF did nothing with it).
To work around this at the moment I nullify the NoteType in the Note in the repository code.
Should I expected AutoMapper to populate all the parent properties with setters or have I missed some configuration? Perhaps there is a glaring flaw in my approach?
When AutoMapper reverses the mapping, it has to collect all information for nested objects from the flat object. Your DTO only carries a value for the mapping NoteType -> NoteTypeDesc. Not for NoteType -> NoteTypeId, so AM really doesn't have any idea where to get that value from.
If you want to rely on flattening only, the only way to change that is to add a flattened NoteTypeId to the DTO besides the unflattened one:
public class NoteDto
{
public int NoteBookId { get; set; }
public short NoteSeqNo { get; set; }
public short NoteTypeId { get; set; } // Not flattened
public short NoteTypeNoteTypeId { get; set; } // Flattened
public string NoteTypeNoteTypeDesc { get; set; }
public string NoteText { get; set; }
}
The alternative is to add this to your mapping:
config.CreateMap<Note, NoteDto>()
.ForMember(dest => dest.NoteTypeId,
e => e.MapFrom(src => src.NoteType.NoteTypeId))
.ReverseMap();
MapFrom-s (including the default unflattening) are reversed now. You can drop ReverseMap and create the maps, ignore Note.NoteType or ignore the offending path, Note.NoteType.NoteTypeDesc.
Sorry for newbie questions, i'm brand new to MVC and OOP
I have the following model for my USER db table
namespace MyApp.Models
{
public class User
{
public int user_id { get; set; }
public string username { get; set; }
public string password { get; set; }
public string salt { get; set; }
public string email { get; set; }
public sbyte status { get; set; }
public System.DateTime creation_date { get; set; }
public sbyte type { get; set; }
public virtual Doctor Doctor { get; set; }
public virtual Owner Owner { get; set; }
public virtual UserToken UserToken { get; set; }
public virtual Veterinarian Veterinarian { get; set; }
}
}
Actually in order to recall a particular USER based on the mail or the id i use a specific class called CustomDbFunctions
namespace MyApp.Models.DAL
{
public static class CustomDbFunctions
{
public static User GetUserEntityFromEmail(string email, DbContext db)
{
return db.Users.FirstOrDefault(u => u.email == (string)email);
}
}
}
in that way i use in my code
User user = CustomDbFunctions.GetUserEntityFromEmail(email, db)
and this it 100% OK with me, but i don't know if this kind of approach is correct or not, or if there's a better way like
//select the single user by calling only the class USER
User mySelectedUser = new User(email)
Thank you very much.
Well for understanding how to access your data in your MVC4 application you could read this tutorial from the Asp.Net MVC main page. Read the whole tutorial about MVC4 and you'll get a solid idea on how to work with it.
But I also recommend this tutorial on a good Entityframework design pattern, it's called Repository Pattern, I just a nice way to get all your code ordered (like all other patterns). Let me know.
I get this webservice error sometimes on a SL5 + EF + WCF app.
"Parameter 'role' of domain operation entry 'AddUserPresentationModelToRole' must be one of the predefined serializable types."
here is a similar error, however his solution doesn't work for me.
I have the codegenned DomainService which surfaces the database entities to my client:
[EnableClientAccess()]
public partial class ClientAppDomainService : LinqToEntitiesDomainService<ClientAppUserEntitlementReviewEntities>
{
public IQueryable<Account> GetAccounts()
{
return this.ObjectContext.Accounts;
}
//..etc...
and my custom service which is surfacing a Presentation model, and db entities.
[EnableClientAccess]
[LinqToEntitiesDomainServiceDescriptionProvider(typeof(ClientAppUserEntitlementReviewEntities))]
public class UserColourService : DomainService
{
[Update(UsingCustomMethod = true)]
public void AddUserPresentationModelToRole(UserPresentationModel userPM, Role role, Reviewer reviewer)
{
...
}
public IDictionary<long, byte> GetColourStatesOfUsers(IEnumerable<RBSUser> listOfUsers, string adLogin)
{
//....
}
}
and the PresentationModel:
public class UserPresentationModel
{
[Key]
public long UserID { get; set; }
public byte UserStatusColour { get; set; }
public string MessageText { get; set; }
[Include]
[Association("asdf", "UserID", "UserID")]
public EntityCollection<Account> Accounts { get; set; }
public DateTime AddedDate { get; set; }
public Nullable<long> CostCentreID { get; set; }
public DateTime? DeletedDate { get; set; }
public string EmailAddress { get; set; }
public long EmployeeID { get; set; }
public string FirstName { get; set; }
public Nullable<bool> IsLeaver { get; set; }
public string LastName { get; set; }
public DateTime LastSeenDate { get; set; }
public string LoginDomain { get; set; }
public string LoginName { get; set; }
public byte WorldBuilderStatusID { get; set; }
}
Also cannot get the solution to reliably fail. It seems whenever I change the service slightly ie make it recompile, everything works.
RIAServices unsupported types on hand-built DomainService - seems to be saying the same thing, that decorating the hand built services with the LinqToEntitiesDomainServiceDescriptionProvider should work.
Possible answer here will post back here too with results.
From Colin Blair:
I am a bit surprised it ever works, I don't think I have seen anyone trying to pass additional entiities into a named update before. It might be a bug in RIA Services that it is working at all. What are you trying to accomplish?
Side note, you have a memory leak with your ObjectContext since it is not getting disposed of correctly. Is there a reason you aren't using the LinqToEntitiesDomainSerivce? It would take care of managing the ObjectContext's lifetime for you.
Results:
1) This makes sense. Have refactored out to more sensible parameters now (ints / strings), and all working.
2) Have brought together my 3 separate services into 1 service, which is using the LinqToEntitiesDomainSerivce. The reason I'd split it out before was the assumption that having a CustomUpdate with a PresentationModel didn't work.. and I had to inherit off DomainService instead. I got around this by making a method:
// need this to avoid compile errors for AddUserPresentationModelToRole.. should never be called
public IQueryable<UserPresentationModel> GetUserPresentationModel()
{
return null;
}
I keep getting error when I try to access a model from an edit or details action.
The model backing the 'InjuriesContext' context has changed since the
database was created. Consider using Code First Migrations to update
the database (http://go.microsoft.com/fwlink/?LinkId=238269).
First I tried adding a migration even though I was sure I hadn't changed anything. Still recieved the same error after an update-database.
Then I removed all the migrations and the database and started a clean database with an inital migration and update. Same error. Nothing was changed.
Model is:
public class InjuriesContext : DbContext
{
public InjuriesContext()
: base("DBCon")
{
}
public DbSet<Patient> Patients { get; set; }
public DbSet<Injury> Injuries { get; set; }
}
public class Injury
{
public int Id { get; set; }
public string Type { get; set; }
public int PatientId { get; set; }
}
Here is controller --
public ActionResult Edit(int id = 0)
{
Injury injury = db.Injuries.Find(id);
if (injury == null)
{
return HttpNotFound();
}
return View(injury);
}
It errors on the injuries.find. I do not have any injuries entered so I expect it to return a 404 like my other controllers but it doesn't like something about this. The only difference between this and my other models is the y to ies for plural. Does Entity Framework not handle this?
There should not be any plural restriction, as you defined everything clearly in your classes anyway.
Have you created the Injuries table?
I belive the table Injury will get created automatically. the variable injury might be a bit close, but I have to test this myself.
Rather try:
public class Injury
{
[Key]
public int Id { get; set; }
[Required]
public string Type { get; set; }
[Required]
public int PatientId { get; set; }
}
private InjuriesContext db = new InjuriesContext();
Injury objInjury = db.Injuries.Find(id);
if (objInjury == null)
{
return HttpNotFound();
}
return View(objInjury);
Hope this helps
It turns out my issue was with multiple contexts. I thought you had to create a separate context for each model class. Apparently Entity Framework needs one context. I went through and created a class for my context and put all my DBsets in that class.
public class ProjContexts : DbContext
{
public ProjContexts()
: base("ProjDBCon")
{
}
public DbSet<Patient> Patients { get; set; }
public DbSet<PreHosp> PreHosps { get; set; }
public DbSet<UserProfile> UserProfiles { get; set; }
public DbSet<Injury> Injuries { get; set; }
}
}
Then I removed all the migrations as per this post and enabled the migrations again did an add migration and update then I got the expected result.
Bottom Line--- Don't have multiple context classes in your project. Not sure if this is possible but after changing the above everything is working as expected. Not sure why it was working when I had two separate contexts and added the third? Maybe because they had foreign keys with one another?
Suppose I have a table:
ID(pk) | HOME_EMAIL | WORK_EMAIL | OTHER_EMAIL
-------------------------------------------------
and the .NET classes
class A {
int id;
List<MyEmail> emails;
}
class MyEmail {
string email;
}
I suppose there's no way to map those (multiple) columns into a single collection in NHibernate, or is there? :)
It's come to a point that we'd rather not tinker with the database schema anymore so we can't do much with the database, if that helps.
I would suggest working with Interfaces so you could do something like this
interface IUser
{
int Id {get; set;}
IEnumerable<string> Emails {get;}
}
class MyUser : IUser
{
public int Id {get; set;}
public IEnumerable<string> Emails
{
get
{
return new [] { SomeEmail, SomeOtherEmail };
}
}
public string SomeEmail { get; set; }
public string SomeOtherEmail { get; set; }
}
Your application can expect an IUser and not care where we got the list of emails. You would map MyUser in NH, while the application does not (and should not) care about the actual implementation.
If it doesn't have to be a collection, but could be a custom type instead, say EmailAddresses which contains three properties:
public class EmailAddresses
{
public virtual string Home { get; set; }
public virtual string Work { get; set; }
public virtual string Other { get; set; }
}
You could use a component to map the three columns into the three properties of this object as a single property on the parent:
public class MyUser
{
...
public virtual EmailAddresses { get; set; }
}
You can map these in NHibernate using components or if you're using Fluent NHibernate with the ComponentMap<T> classmap (automapper can't do components).
There is a feature that's very close to what you want, <dynamic-component>
The documentation at http://nhibernate.info/doc/nh/en/index.html#components-dynamic should get you started.