I am creating a example for better understanding.
[CustomValidator("Property1","Property2", ErrorMessage= "Error1")]
[CustomValidator("Property3","Property4", ErrorMessage= "Error1")]
public class MyViewModel
{
public string Property1 {get; set;}
public string Property2 {get; set;}
public string Property3 {get; set;}
public string Property4 {get; set;}
}
[AttributeUsage(AttributeTargets.Class, AllowMultiple = true, Inherited = false)]
public class CustomValidator : ValidationAttribute
{
All the required stuff is written.
}
Only the second validator (or the last one in the list) gets fired and ignores the first one.
I am not sure if this is the right approach for this scenario.
Any suggestions?
if you are using Linq to SQL why not try something like this
add a rule violations class to handle rule violations
public class RuleViolation
{
public string ErrorMessage { get; private set; }
public string PropertyName { get; private set; }
public RuleViolation(string errorMessage)
{
ErrorMessage = errorMessage;
}
public RuleViolation(string errorMessage, string propertyName)
{
ErrorMessage = errorMessage;
PropertyName = propertyName;
}
}
now on your data class
[Bind(Exclude="ID")]
public partial class Something
{
public bool IsValid
{
get { return (GetRuleViolations().Count() == 0); }
}
public IEnumerable<RuleViolation> GetRuleViolations()
{
if (String.IsNullOrEmpty(Name.Trim()))
yield return new RuleViolation("Name Required", "Name");
if (String.IsNullOrEmpty(LocationID.ToString().Trim()))
yield return new RuleViolation("Location Required", "LocationID");
yield break;
}
partial void OnValidate(ChangeAction action)
{
if (!IsValid)
throw new ApplicationException("Rule violations prevent saving");
}
}
and in your controller's methods for updating, use the updatemodel method for changing properties
Something something = somethingRepo.GetSomething(id);
try
{
//update something
UpdateModel(something);
somethingRepo.Save();
return RedirectToAction("Index");
}
catch
{
ModelState.AddRuleViolations(something.GetRuleViolations());
return View(something);
}
this way you can just add rules to your data class as it changes and it will be reflected in your updates, etc
I found another question that answers this. You have to override Attribute.TypeId.
Custom validation attribute with multiple instances problem
you dont really need all that code use data annotations by creating a metaData class for your model link text
that should set you on the right road also read up on html helpers and buddy classes ( this is what they call em)
I had the same problem.
I come up with the following solution.
Your POCO class might implement interface IValidatableObject.
This requires from you to implement the following method.
public virtual IEnumerable<ValidationResult> Validate(ValidationContext validationContext)
{
// return list of appropriate ValidationResult object
var customResult = new List<ValidationResult>();
customResult.Add(new ValidationResult("message", new List<string>(){"Property1"});
}
You can place any validation logic there. This has also the advantage over class-level attribute. Class-level attribute can only be shown in ValidationSummary (they are not related to any property). In contrast to it you can set specific members while returning ValidationResult. This allows to show validation information beside specific control to which the message concerns.
Related
the class has a property UserId which type of ApplicationUser and its required, and of course we can not pass it from view to controller because of security reasons.
Now when the controller checks the state of model the model is not in a correct state because there is no UserId Value and it returns the view back, if I use the view model the class has more than 50 properties and assigning the values from view model to class and then save it it is very tedious and difficult to do it for saving editing and so on, any advice to come out from this problem
thanks
You can use Action Filters attribute to Auto Bind Property.
Let's say we have UserSettingMetaModel with UserId property.
public interface IAutoBindingUserId
{
int UserId { get; set; }
}
public class UserSettingMetaModel : IAutoBindingUserId
{
public int Id { get; set; }
[Required]
public int UserId { get; set; }
// The rest of properties
}
In UserSetting controller, we have
[HttpPost]
[AutoBindProperty]
public JsonResult Add(UserSettingMetaModel metaModel)
{
if (ModelState.IsValid)
{
// do something
}
}
Behind the sense of AutoBindProperty Attribute
[AttributeUsage(AttributeTargets.Method)]
public sealed class AutoBindPropertyAttribute : ActionFilterAttribute
{
public override void OnActionExecuting(ActionExecutingContext filterContext)
{
var actionParams = filterContext.ActionParameters.Values.FirstOrDefault();
Bind<IAutoBindingUserId>(model => model.UserId = 123, actionParams);
// Assuming that you know the way get userId value here.
}
private static void Bind<T>(Action<T> doBinding, object actionParams) where T : class
{
if (actionParams is T model)
{
doBinding(model);
}
}
}
anyhow thanks for the cooperation for solving this problem
the easiest way that i found to return a UserId from a class, just one line of code is required inside class constructor
public constructor()
{
UserId = ClaimsPrincipal.Current.Identity.GetUserId();
}
I am building a simple search, sort, page feature. I have attached the code below.
Below are the usecases:
My goal is to pass the "current filters" via each request to persist them particularly while sorting and paging.
Instead of polluting my action method with many (if not too many) parameters, I am thinking to use a generic type parameter that holds the current filters.
I need a custom model binder that can be able to achieve this.
Could someone please post an example implementation?
PS: I am also exploring alternatives as opposed to passing back and forth the complex objects. But i would need to take this route as a last resort and i could not find a good example of custom model binding generic type parameters. Any pointers to such examples can also help. Thanks!.
public async Task<IActionResult> Index(SearchSortPage<ProductSearchParamsVm> currentFilters, string sortField, int? page)
{
var currentSort = currentFilters.Sort;
// pass the current sort and sortField to determine the new sort & direction
currentFilters.Sort = SortUtility.DetermineSortAndDirection(sortField, currentSort);
currentFilters.Page = page ?? 1;
ViewData["CurrentFilters"] = currentFilters;
var bm = await ProductsProcessor.GetPaginatedAsync(currentFilters);
var vm = AutoMapper.Map<PaginatedResult<ProductBm>, PaginatedResult<ProductVm>>(bm);
return View(vm);
}
public class SearchSortPage<T> where T : class
{
public T Search { get; set; }
public Sort Sort { get; set; }
public Nullable<int> Page { get; set; }
}
public class Sort
{
public string Field { get; set; }
public string Direction { get; set; }
}
public class ProductSearchParamsVm
{
public string ProductTitle { get; set; }
public string ProductCategory { get; set; }
public Nullable<DateTime> DateSent { get; set; }
}
First create the Model Binder which should be implementing the interface IModelBinder
SearchSortPageModelBinder.cs
public class SearchSortPageModelBinder<T> : IModelBinder
{
public Task BindModelAsync(ModelBindingContext bindingContext)
{
if (bindingContext == null)
{
throw new ArgumentNullException(nameof(bindingContext));
}
SearchSortPage<T> ssp = new SearchSortPage<T>();
//TODO: Setup the SearchSortPage<T> model
bindingContext.Result = ModelBindingResult.Success(ssp);
return TaskCache.CompletedTask;
}
}
And then create the Model Binder Provider which should be implementing the interface IModelBinderProvider
SearchSortPageModelBinderProvider.cs
public class SearchSortPageModelBinderProvider : IModelBinderProvider
{
public IModelBinder GetBinder(ModelBinderProviderContext context)
{
if (context == null)
{
throw new ArgumentNullException(nameof(context));
}
if (context.Metadata.ModelType.GetTypeInfo().IsGenericType &&
context.Metadata.ModelType.GetGenericTypeDefinition() == typeof(SearchSortPage<>))
{
Type[] types = context.Metadata.ModelType.GetGenericArguments();
Type o = typeof(SearchSortPageModelBinder<>).MakeGenericType(types);
return (IModelBinder)Activator.CreateInstance(o);
}
return null;
}
}
And the last thing is register the Model Binder Provider, it should be done in your Startup.cs
public void ConfigureServices(IServiceCollection services)
{
.
.
services.AddMvc(options=>
{
options.ModelBinderProviders.Insert(0, new SearchSortPageModelBinderProvider());
});
.
.
}
I'm starting to working on ASP.NET using MVC. I writing to action results, one of them is a HTTP GET and the another HTTP POST
[HttpGet]
public ActionResult DoTest()
{
Worksheet worksheets = new worksheets(..);
return View(w);
}
[HttpPost]
public ActionResult DoTest(Worksheet worksheet)
{
return PartialView("_Problems", worksheet);
}
Now, Worksheet class has a property called Problems and this is a collection, but uses as an abstract class item.
public class Worksheet
{
public List<Problem> Problems { get; set; }
}
Here's my abstract class and one implementation
public abstract class Problem
{
public Problem() { }
public int Id { get; set; }
public abstract bool IsComplete { get; }
protected abstract bool CheckResponse();
}
public class Problem1 : Problem
{
...
public decimal CorrectResult { get; set; }
// this is the only property of my implementation class which I need
public decimal? Result { get; set;}
public override bool IsComplete
{
get { return Result.HasValue; }
}
protected override bool CheckResponse()
{
return this.CorrectResult.Equals(this.Result.Value);
}
}
I have right now, many implementations of Problem class, but I really need to get just one value of my implementation class. But it thrown the above image error.
What can I do to allow model binder recover that part of my abstracts classes
The following code would not compile:
var problem = new Problem();
... because the Problem class is abstract. The MVC engine cannot just create a Problem directly. Unless you give it some way to know which type of Problem to instantiate, there's nothing it can do.
It is possible to create your own ModelBinder implementation, and tell MVC to use it. Your implementation could be tied to a Dependency Injection framework, for example, so that it knows to create a Problem1 whenever a Problem class is requested.
Or you could simply change your action method to take a concrete type:
public ActionResult DoTest(IEnumerable<Problem1> problems)
{
return PartialView("_Problems",
new Worksheet {
Problems = problems.Cast<Problem>().ToList()
});
}
Building on Ladislav's answer to
Entity Framework Code First and Collections of Primitive Types
I'm attempting to create a wrapper type EfObservableCollection<T> around an ObservableCollection<T> that has an additional helper property to simplify persistence (certainly this solution has trade-offs, but it's seems workable for my domain).
However, properties of type EfObservableCollection<T> seem to be ignored by EF. No appropriate columns are created in the database. Guessing that implementing IEnumerable<T> might trigger EF to ignore that type, I commented out that implementation with no change in behavior.
What am I missing here?
Entity Class
public class A
{
[DataMember]
public long Id { get; set; }
[DataMember]
public string Text { get; set; }
// Tags is not persisted
[DataMember]
public EfObservableCollection<string> Tags { get; set; }
}
Wrapper Class
[ComplexType]
public class EfObservableCollection<T> : IEnumerable<T>
{
const string VALUE_SEPARATOR = "\x8"; // Backspace character. Unlikely to be actually entered by a user. Assumes ASCII or UTF-8.
readonly string[] VALUE_SEPARATORS = new string[] { VALUE_SEPARATOR };
[NotMapped]
protected ObservableCollection<T> Collection { get; private set; }
public EfObservableCollection()
{
Collection = new ObservableCollection<T>();
}
[DataMember]
public string PersistHelper
{
get
{
string serializedValue = string.Join(VALUE_SEPARATOR, Collection.ToArray());
return serializedValue;
}
set
{
Collection.Clear();
string[] serializedValues = value.Split(VALUE_SEPARATORS, StringSplitOptions.None);
foreach (string serializedValue in serializedValues)
{
Collection.Add((T)Convert.ChangeType(serializedValue, typeof(T))); // T must implement IConvertable, else a runtime exception.
}
}
}
public void Add(T item)
{
Collection.Add(item);
}
IEnumerator<T> GetEnumerator()
{
return Collection.GetEnumerator();
}
System.Collections.IEnumerator System.Collections.IEnumerable.GetEnumerator()
{
return GetEnumerator();
}
}
It turns out that Entity Framework does not like the generic class EfObservableCollection<T>.
If I derive a non-generic class from that class, data is persisted as expected:
[ComplexType]
public class EfObservableCollectionString : EfObservableCollection<string>
{
}
Joining backspace with list of strings causes cleaning last character in each string item.
I think serialization to json using System.Web.Script.Serialization.JavaScriptSerializer is better.
I have a layered application that send commands to the business layer (actually, the application is based on ncqrs framework, but I don't think it's important here).
A command looks like this :
public class RegisterUserCommand : CommandBase
{
public string UserName { get; set; }
public string Email{ get; set; }
public DateTime RegistrationDate { get; set; }
public string ApiKey {get; set;} // edit
}
There is no logic in this class, only data.
I want to have the users type their user name, email and I want the system to use the current date to build the command.
What is best between :
create a strongly typed view based on the RegisterUserCommand, then inject the date and the APi Key just before sending it to the business layer ?
create a RegisterUserViewModel class, create the view with this class and create the command object based on the view input ?
I wrote the following code (for the solution n°2) :
public class RegisterController : Controller
{
//
// GET: /Register/
public ActionResult Index()
{
return View();
}
[HttpPost]
public ActionResult Index(RegisterUserViewModel registrationData)
{
var service = NcqrsEnvironment.Get<ICommandService>();
service.Execute(
new RegisterUserCommand
{
RegistrationDate = DateTime.UtcNow,
Email= registrationData.Email,
UserName= registrationData.Name,
ApiKey = "KeyFromConfigSpecificToCaller" // edit
}
);
return View();
}
public class RegisterUserViewModel
{
[Required]
[StringLength(16)]
public string Name { get; set; }
[Required]
[StringLength(64)]
public string Email{ get; set; }
}
}
This code is working... but I wonder if I choose the correct way...
thanks for advises
[Edit] As the Datetime seems to cause misunderstanding, I added another property "ApiKey", that should also be set server side, from the web layer (not from the command layer)
[Edit 2] try the Erik suggestion and implement the 1st solution I imagined :
[HttpPost]
public ActionResult Index(RegisterUserCommand registrationCommand)
{
var service = NcqrsEnvironment.Get<ICommandService>();
registrationCommand.RegistrationDate = DateTime.UtcNow;
registrationCommand.ApiKey = "KeyFromConfigSpecificToCaller";
service.Execute(
registrationCommand
);
return View();
}
... Is it acceptable ?
I think you would be better off with option #2, where you would have a separate ViewModel and a Command. While it may seem redundant (to an extent), your commands are really messages from your web server to your command handler. Those messages may not be formatted the same as your ViewModel, nor should they be. And if you're using NCQRS as is, you would then have to map your commands to your AR methods and constructors.
While it may save you a little bit of time, I think you pigeon-hole yourself in to modeling your domain after your ViewModels, and that should not be the case. Your ViewModels should be a reflection of what your user experiences and sees; your domain should be a reflection of your business rules and knowledge, and are not always reflected in your view.
It may seem like a bit more work now, but do yourself a favor and keep your commands separate from your view models. You'll thank yourself later.
I hope this helps. Good luck!
I would recommend putting this into the constructor of the RegisterUserCommand class. That way the default behavior is always to set it to DateTime.UtcNow, and if you need to set it to something explicitly you can just add it to the object initializer. This will also help in scenarios where you're using this class in other parts of your project, and you forget to set the RegistrationDate explicitly.
public class RegisterUserCommand : CommandBase
{
public string UserName { get; set; }
public string Email{ get; set; }
public DateTime RegistrationDate { get; set; }
public RegisterUserCommand()
{
RegistrationDate = DateTime.UtcNow;
}
}
And the Controller
public class RegisterController : Controller
{
//
// GET: /Register/
public ActionResult Index()
{
return View();
}
[HttpPost]
public ActionResult Index(RegisterUserViewModel registrationData)
{
var service = NcqrsEnvironment.Get<ICommandService>();
service.Execute(
new RegisterUserCommand
{
Email= registrationData.Email,
OpenIdIdentifier = registrationData.OpenIdIdentifier
}
);
return View();
}
public class RegisterUserViewModel
{
[Required]
[StringLength(16)]
public string Name { get; set; }
[Required]
[StringLength(64)]
public string Email{ get; set; }
}
}
I would use number 1 and use the system.componentmodel.dataannotations.metadatatype for validation.
I created an example (answer) for another SO question Here.
This allows you to keep your model in another library, validate the fields and show the fields like you would internal/private classes with DataAnnotations. I'm not a big fan of creating a completely separate class for a view that has no additional value while having to ORM the data back to another class. (If you had additional values like dropdown list values, or default values then I think it would make sense).
Instead of
[HttpPost]
public ActionResult Index(RegisterUserViewModel registrationData)
{
var service = NcqrsEnvironment.Get<ICommandService>();
service.Execute(
new RegisterUserCommand
{
RegistrationDate = DateTime.UtcNow,
Email= registrationData.Email,
UserName= registrationData.Name,
ApiKey = "KeyFromConfigSpecificToCaller" // edit
}
);
return View();
}
You can have
[HttpPost]
public ActionResult Index(RegisterUserCommand registrationData)
{
var service = NcqrsEnvironment.Get<ICommandService>();
registrationData.ApiKey = "KeyFromConfigSpecificToCaller";
service.Execute(registrationData);
return View();
}