I have following entities:
abstract class User
{
string Id
string Name
}
class UserA: User
{
string PropA
}
class UserB : User
{
string PropB
}
It is a good solution to have a unique create (post) with a dynamic parameter and instantiate the subclasses according to a property?
[HttpPost]
public IActionResult Create([FromBody]dynamic data)
{
if (data.PROP == null)
{
_context.Users.Add(new UserA(data.PropA));
}
else
{
_context.Users.Add(new UserB(data.PropB));
}
...
Don't use dynamic. I'm actually kind of surprised that works at all. Though there's no indication that you've actually tested this code yet, so perhaps it doesn't. The modelbinder needs to know a concrete type to bind to, so that it can determine how to map the values onto the destination instance. Without strong types, it can't do anything but make everything a string, since that is how it comes in the request body.
Anyways, for something like this, the correct approach is to use a view model. Your view model should contain all the properties for all the various possible derived types. Again, the modelbinder needs these to determine how to map the data from the request body over, so if a property doesn't exist, it will simply discard the associated data.
This is also why you cannot simply use the base class. If this were a normal method, you could do something like:
public IActionResult Create([FromBody]User data)
Then, inside, you could use pattern matching or similar to cast to the correct derived type. This works because ultimately, the object in memory would actually be an instance of something like UserA, and you're simply up-casting it to User. As a result, you can always cast it back to UserA. However, actions are different. What's coming in from the request is not an object instance. The modelbinder serves to create an object instance out of it, by inspecting the parameter it needs to bind to. If that parameter is of type User, then it will fill the properties on User, and discard everything else. As a result, the object in memory is just User, and there's no way to cast to something like UserA - at least in terms of having all the values that were actually posted for an instance of UserA being on the object.
Which brings us back to the view model:
public class UserViewModel
{
public string Id { get; set; }
public string Name { get; set; }
public string PropA { get; set; }
public string PropB { get; set; }
}
Then, have your action accept that as a param:
public IActionResult Create([FromBody]UserViewModel data)
Then, inside:
if (!string.IsNullOrWhiteSpace(data.PropA))
{
// UserA was posted, map data to an instance of UserA
}
Similarly for UserB. If you like, you could also post an explicit "type" along with the data and switch on that to instantiate the right type. It's up to you. To reduce code duplication, you can instantiate the right type, but store it in an variable of type User. Then, if you need to get back at the correct type, you can use pattern matching:
User user;
switch (data.Type)
{
case "UserA":
user = new UserA
{
Id = data.Id,
Name = data.Name,
PropA = data.PropA
};
break;
// etc.
default:
user = new User
{
Id = data.Id,
Name = data.Name
};
break;
}
Then later:
switch (user)
{
case UserA userA:
// do something specific with `userA`
// etc.
}
Or:
if (user is UserA userA)
{
// do something with `userA`
}
Related
I'm looking for a way to intercept Entity Framework's lazy load proxy implementation, or otherwise control what is returned when accessing a Navigation property that may have no value in the database.
An example of what I'm looking for is this Contact class with mailing address, business phone, etc. that may or may not have a contact person.
public partial class Contact
{
private Nullable<System.Guid> _personId;
public Nullable<System.Guid> PersonId
{
get { return _personId; }
set { SetProperty(ref _personId, value); }
}
public virtual Person Person{ get; set; }
// mailing address, other properties...
}
public partial class Person
{
private string _firstName;
public string FirstName
{
get { return _firstName; }
set { SetProperty(ref _firstName, value); }
}
private string _lastName;
public string LastName
{
get { return _lastName;}
set { SetProperty(ref _lastName;value); }
}
}
It is very useful in ASP.net Razor pages, WPF or ad-hoc reporting tools, to be able to use expressions like:
Contact c = repo.GetContact(id);
Console.WriteLine("Contact Person " + c.Person.FirstName);
Which of course fails if there is no PersonId, and hence contact.Person is null.
Tools like Ideablade Devforce have a mechanism to return a "NullEntity" for Person in this case, which allows the WriteLine to succeed. Additionally, the NullEntity for Person can be configured to have a sensible value for FirstName, like "NA".
Is there some way to override the Dynamic Proxy mechanism in EF, or otherwise intercept the reference to Person from Contact to enable this scenario?
I have investigated IDbCommandInterceptor, but that does not seem to intercept virtual navigation to individual entity properties, only navigation to entity collections.
Update _____________________________________
To elaborate on my original question, I can't modify the expression by introducing null conditional operators into the them, as these expressions are incorporated into WPF, ASP.Net Razor binding expressions, and/or report data fields, created by other developers or authors. Also, there may be multiple layers of null properties to deal with, e.g. Contact.Person.Spouse.FirstName, where either Person and/or Spouse might be a "null" property. The Devforce Ideablade implementation deals with this perfectly, but is unfortunately not an option on my current project.
you can use a null-conditional operator from c# like this
c.Person?.FirstName
This means that when Person == null , return null or otherwise return FirstName. You would still need to handle the null value
See : https://learn.microsoft.com/en-us/dotnet/csharp/language-reference/operators/member-access-operators#null-conditional-operators--and-
I have a one to many relationship as outlined below. In some parts of the business layer there are queries of the Item table and in others the Client table (as well as its Items). LazyLoading and ProxyCreation are both false, reference loop handling is set to ignore.
public class Client {
public virtual ICollection<Item> Items { get; set; }
public string Name {get;set;}
}
public class Item {
public virtual Client TheClient {get;set;}
public string ItemProp {get;set;}
// another 10 properties or so
}
myitems = dbContextScopeX.Items.Include(x => x.Client).ToList();
The view has a list of items with the need to show the Client's Name (in my example). I am looking for item.Client.Name ultimate, however when myitems gets queries/serialized it contains:
myitems.Client.Items
If I set the attribute [JsonIgnore] on the Client's Item property it never comes through the graph which I need it to in other places. Is there a way to get myItems.Client.Name without having to get myitems.Client.Items in the query or without having to create an anonymous projection for the Item array?
Project the Item properties you want (be they simple or complex type) along with just the Client name into an anonymous type and serialize that.
myitems = dbContextScopeX.Items.Include(x => x.Client)
.Select(i=>new {
ItemProp = i.ItemProp,
ItemCollection = i.ItemCollection,
...
ClientName = i.Client.Name
}).ToList();
Only caveat is you have to do some manual work if you want to deserialize this back into entities.
What's the best way to set default properties for new entities in DDD? Also, what's the best way to set default states for complex properties (eg. collections)?
My feeling is that default values should be in the models themselves as they are a form of business rule ("by default, we want X's to be Y & Z"), and the domain represents the business. With this approach, maybe a static "GetNew()" method on the model itself would work:
public class Person {
public string Name { get; set; }
public DateTime DateOfBirth { get; set; }
public bool IsAlive { get; set; }
public List Limbs { get; set; }
public static Person GetNew() {
return new Person() {
IsAlive = true,
Limbs = new List() { RightArm, LeftArm, RightLeg, LeftLeg }
}
}
}
Unfortunately in our case, we need the collection property to be set to all members of another list, and as this model is decoupled from its Repository/DbContext it doesn't have any way of loading them all.
Crappy solution would be to pass as parameter :
public static Person GetNew(List<Limb> allLimbs) {
return new Person() {
IsAlive = true,
Limbs = allLimbs
}
}
Alternatively is there some better way of setting default values for simple & complex model properties?
This is an instance of the factory pattern in DDD. It can either be a dedicated class, such as PersonFactory, or a static method, as in your example. I prefer the static method because I see no need to create a whole new class.
As far as initializing the collection, the GetNew method with the collection as a parameter is something I would go with. It states an important constraint - to create a new person entity you need that collection. The collection instance would be provided by an application service hosting the specific use case where it is needed. More generally, default values could be stored in the database, in which case the application service would call out to a repository to obtain the required values.
Take a look at the Static Builder in Joshua Bloch's Effective Java (Second Edition). In there, you have a static builder class and you chain calls to set properties before construction so it solves the problem of either having a constructor that takes a ton of arguments or having to put setters on every property (in which case, you effectively have a Struct).
When using code-first EntityFramework, I need one property to be set before the others - how do I specify the order that it calls the property sets, when it is creating the objects from the database?
E.g.
public class Person
{
public string Name { get; set; }
public string Something
{
get { return something; }
set
{
something = value + " for " + Name;
}
}
private string something;
}
In the code above, I need the Name property to already have been set by the time it sets the Something property.
This isn't the actual example - I know there are other ways to achieve that literally, but I'm not after those, just how I can tell EF to set Name before Something.
I am trying to understand the context of your question. I am going to make the assumptions that:
The value passed to the setter is not another calculated property
the value passed to the setter is intended to be stored in the database
If name is updated you would want Something to be updated to reflect the new name?
I think your mistake here is trying to add a derived portion to the value you are looking to store. Calculate the pretty name in another property, or on a get:
UPDATE had an example overriding the get on the Something Property, but removed as I feel it is bad practice.
public class Person
{
public string Name { get; set; }
public string Something { get; set; }
public string getFancySomething {
{ return Something + " for " + Name; }
}
}
Finally - (and here is where I could use some help as I have not run into a situation where I needed to do something like this) my guess is that you do not need to be storing the partially calculated property Something in the way you were attempting, but if you do need to, I think additional detail might help someone provide you with a better answer.
UPDATE 2
As described in my comments - not sure this would work, and it feels very wrong, but you could try something like:
modelBuilder.Entity<Person>().Ignore(x => x.Something);
and then in the setter:
public class Person
{
public string Name {
get { return Name; }
set {
Name = value;
Something = lookup();
}
}
}
Again this will depend on you needs, and would not satisfy any need to pass this value in, and I am not sure this is a great idea.
i have a question.i have a method (Filter),i want to pass T dynamic.but it dosen`t accept.how can i do it?
public List<T> Filter<T>(string TypeOfCompare)
{
List<T> ReturnList2 = new List<T>();
return ReturnList2;
}
IList MakeListOfType(Type listType)
{
Type listType1 = typeof(List<>);
Type specificListType = listType.MakeGenericType(listType1);
return (IList)Activator.CreateInstance(specificListType);
}
Filter < ConstructGenericList(h) > ("s");
IList MakeListOfType(Type listType)
{
Type listType1 = typeof(List<>);
Type specificListType = listType.MakeGenericType(listType1);
return (IList)Activator.CreateInstance(specificListType);
}
It should be the other way round, you should call MakeGenericType on the generic type definition, not on the generic type argument. So the code becomes this:
IList MakeListOfType(Type elementType)
{
Type listType = typeof(List<>);
Type specificListType = listType.MakeGenericType(elementType);
return (IList)Activator.CreateInstance(specificListType);
}
(note that I changed the variables names to make the code clearer)
Generic parameters must have a type able to be determined at compile time (without resorting to something like functional type inference that some other languages have). So, you can't just stick a function between the angle brackets to get the type you want.
Edit:
Now that I know what you're trying to do, I would suggest a different approach entirely.
You mention that you are using Entity Framework, and you are trying to use one method to get a list of different types of objects. Those objects -- like Student and Teacher -- must have something in common, though, else you would not be trying to use the same method to retrieve a list of them. For example, you may just be wanting to display a name and have an ID to use as a key.
In that case, I would suggest defining an interface that has the properties common to Student, Teacher, etc. that you actually need, then returning a list of that interface type. Within the method, you would essentially be using a variant of the factory pattern.
So, you could define an interface like:
public interface INamedPerson
{
int ID { get; }
string FirstName { get; }
string LastName { get; }
}
Make your entities implement this interface. Auto-generated entities are (typically) partial classes, so in your own, new code files (not in the auto-generated code files themselves), you would do something like:
public partial class Student : INamedPerson
{
public int ID
{
get
{
return StudentId;
}
}
}
and
public partial class Teacher : INamedPerson
{
public int ID
{
get
{
return TeacherId;
}
}
}
Now, you may not even need to add the ID property if you already have it. However, if the identity property in each class is different, this adapter can be one way to implement the interface you need.
Then, for the method itself, an example would be:
public List<INamedPerson> MakeListOfType(Type type)
{
if (type == typeof(Student))
{
// Get your list of students. I'll just use a made-up
// method that returns List<Student>.
return GetStudentList().Select<Student, INamedPerson>(s => (INamedPerson)s)
.ToList<INamedPerson>();
}
if (type == typeof(Teacher))
{
return GetTeacherList().Select<Teacher, INamedPerson>(t => (INamedPerson)t)
.ToList<INamedPerson>();
}
throw new ArgumentException("Invalid type.");
}
Now, there are certainly ways to refine this pattern. If you have a lot of related classes, you may want to use some sort of dependency injection framework. Also, you may notice that there is a lot of duplication of code. You could instead pass a function (like GetStudentList or GetTeacherList) by doing something like
public List<INamedPerson> GetListFromFunction<T>(Func<IEnumerable<T>> theFunction) where T : INamedPerson
{
return theFunction().Select<T, INamedPerson>(t => (INamedPerson)t).ToList<INamedPerson>();
}
Of course, using this function requires, once again, the type passed in to be known at compile time. However, at some point, you're going to have to decide on a type, so maybe that is the appropriate time. Further, you can make your life a little simpler by leaving off the generic type at method call time; as long as you are passing in a function that takes no arguments and returns an IEnumerable of objects of the same type that implement INamedPerson, the compiler can figure out what to use for the generic type T.