How to test whether a lazy loaded child collection has been loaded - entity-framework-6

Suppose I have users with tags being lazy loaded:
public class User
{
public int UserId { get; set; }
public virtual ICollection<Tag> Tags { get; set; }
}
If I get a user this way, I know that the tags aren't loaded:
User myUser;
using (var context = new MyContext())
{
myUser = context.Users.Find(4);
}
How do I test the Tags collection presence outside of the using clause?
if (myUser.Tags == null) // throws an ObjectDisposedException
I could use a try/catch but there must be a better way.

The only way I can think of is to be able to do a non virtual call (similar to when you do base.Something in a derived class) to the class property getter. Since there is no way to do that with pure C# or reflection, I've ended up with the following helper method which utilizes System.Reflection.Emit LCG (lightweight code generation) to emit Call IL instruction instead of the normal Callvirt:
using System;
using System.Linq.Expressions;
using System.Reflection;
using System.Reflection.Emit;
public static class Utils
{
public static TValue GetClassValue<TSource, TValue>(this TSource source, Expression<Func<TSource, TValue>> selector)
where TSource : class
{
Func<TSource, TValue> getValue = null;
if (source.GetType() != typeof(TSource))
{
var propertyAccessor = selector.Body as MemberExpression;
if (propertyAccessor != null)
{
var propertyInfo = propertyAccessor.Member as PropertyInfo;
if (propertyInfo != null)
{
var getMethod = propertyInfo.GetGetMethod();
if (getMethod != null && getMethod.IsVirtual)
{
var dynamicMethod = new DynamicMethod("", typeof(TValue), new[] { typeof(TSource) }, typeof(Utils), true);
var il = dynamicMethod.GetILGenerator();
il.Emit(OpCodes.Ldarg_0);
il.EmitCall(OpCodes.Call, getMethod, null);
il.Emit(OpCodes.Ret);
getValue = (Func<TSource, TValue>)dynamicMethod.CreateDelegate(typeof(Func<TSource, TValue>));
}
}
}
}
if (getValue == null)
getValue = selector.Compile();
return getValue(source);
}
}
It can be used for both single and collection type navigation properties like this:
if (myUser.GetClassValue(x => x.Tags) == null)

Another solution using a naive try/catch, not sure about the performance compared to the other answer:
using System;
public static class EntityFrameworkExtensions
{
public static bool IsCollectionLoaded<TSource, TValue>(this TSource source, Func<TSource, TValue> selector)
where TSource : class
{
try
{
return (selector(source) != null);
}
catch (ObjectDisposedException)
{
return false;
}
}
}
usage:
if (myUser.IsCollectionLoaded(x => x.Tags))

Related

OData Client Request Pipeline not working in Odata V4

I am following the sample blog below to remove and add properties in request ODataEntry class.
http://blogs.msdn.com/b/odatateam/archive/2013/07/26/using-the-new-client-hooks-in-wcf-data-services-client.aspx
But even if the code works fine and adds and removes the properties correctly when I put breakpoint, all the entity properties goes to server un changed.
Only difference I see this I am using the OData V4 and new Ondata client to hook up.
My code looks below.
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;
using Client.Default;
namespace Client
{
using Client.MvcApplication1.Models;
using Microsoft.OData.Core;
internal class Program
{
private static void Main(string[] args)
{
Container container = new Container(new Uri("http://localhost:55000/api/"));
container.Configurations.RequestPipeline.OnEntryEnding(
w =>
{
w.Entry.RemoveProperties("Name");
});
Test test = new Test();
test.Name = "Foo";
CustomFields cs = new CustomFields { ServiceId = 3 };
cs.Foo1 = 2;
test.S_1 = cs;
container.AddToTests(test);
container.SaveChanges();
}
}
public static class Extensions
{
public static void RemoveProperties(this ODataEntry entry, params string[] propertyNames)
{
var properties = entry.Properties as List<ODataProperty>;
if (properties == null)
{
properties = new List<ODataProperty>(entry.Properties);
}
var propertiesToRemove = properties.Where(p => propertyNames.Any(rp => rp == p.Name));
foreach (var propertyToRemove in propertiesToRemove.ToArray())
{
properties.Remove(propertyToRemove);
}
entry.Properties = properties;
}
public static void AddProperties(this ODataEntry entry, params ODataProperty[] newProperties)
{
var properties = entry.Properties as List<ODataProperty>;
if (properties == null)
{
properties = new List<ODataProperty>(entry.Properties);
}
properties.AddRange(newProperties);
entry.Properties = properties;
}
}
}
If I change and start listening to RequestPipeline.OnEntryStarting I get the validation error that new property is not defined in owning entity. But as per code for Microsoft.OData.CLient this error should not occure as there is a check for IEdmStructuredType.IsOpen but still error occurs. So issue seems deep in how owningStructuredType is calculated. On my container I do see the correct edm model with entities marked as IsOpen = true.
Odata lib code which should pass but is failing
internal static IEdmProperty ValidatePropertyDefined(string propertyName, IEdmStructuredType owningStructuredType)
{
Debug.Assert(!string.IsNullOrEmpty(propertyName), "!string.IsNullOrEmpty(propertyName)");
if (owningStructuredType == null)
{
return null;
}
IEdmProperty property = owningStructuredType.FindProperty(propertyName);
// verify that the property is declared if the type is not an open type.
if (!owningStructuredType.IsOpen && property == null)
{
throw new ODataException(Strings.ValidationUtils_PropertyDoesNotExistOnType(propertyName, owningStructuredType.ODataFullName()));
}
return property;
}
Client code:
container.Configurations.RequestPipeline.OnEntryStarting(
w =>
{
w.Entry.RemoveProperties("Name");
w.Entry.AddProperties(new ODataProperty
{
Name = "NewProperty",
Value = 1
});
});
Error:
The property 'NewProperty' does not exist on type 'Client.MvcApplication1.Models.Test'. Make sure to only use property names that are defined by the type.
at Microsoft.OData.Core.WriterValidationUtils.ValidatePropertyDefined(String propertyName, IEdmStructuredType owningStructuredType)
at Microsoft.OData.Core.JsonLight.ODataJsonLightPropertySerializer.WriteProperty(ODataProperty property, IEdmStructuredType owningType, Boolean isTopLevel, Boolean allowStreamProperty, DuplicatePropertyNamesChecker duplicatePropertyNamesChecker, ProjectedPropertiesAnnotation projectedProperties)
at Microsoft.OData.Core.JsonLight.ODataJsonLightPropertySerializer.WriteProperties(IEdmStructuredType owningType, IEnumerable`1 properties, Boolean isComplexValue, DuplicatePropertyNamesChecker duplicatePropertyNamesChecker, ProjectedPropertiesAnnotation projectedProperties)
at Microsoft.OData.Core.JsonLight.ODataJsonLightWriter.StartEntry(ODataEntry entry)
at Microsoft.OData.Core.ODataWriterCore.<>c__DisplayClass14.<WriteStartEntryImplementation>b__12()
at Microsoft.OData.Core.ODataWriterCore.InterceptException(Action action)
at Microsoft.OData.Core.ODataWriterCore.WriteStartEntryImplementation(ODataEntry entry)
at Microsoft.OData.Core.ODataWriterCore.WriteStart(ODataEntry entry)
at Microsoft.OData.Client.ODataWriterWrapper.WriteStart(ODataEntry entry, Object entity)
at Microsoft.OData.Client.Serializer.WriteEntry(EntityDescriptor entityDescriptor, IEnumerable`1 relatedLinks, ODataRequestMessageWrapper requestMessage)
at Microsoft.OData.Client.BaseSaveResult.CreateRequestData(EntityDescriptor entityDescriptor, ODataRequestMessageWrapper requestMessage)
at Microsoft.OData.Client.BaseSaveResult.CreateChangeData(Int32 index, ODataRequestMessageWrapper requestMessage)
at Microsoft.OData.Client.SaveResult.CreateNonBatchChangeData(Int32 index, ODataRequestMessageWrapper requestMessage)
at Microsoft.OData.Client.SaveResult.CreateNextChange()
I use partial classes defined on the client to add the extra properties that I need there. This allows me to put any logic in them as well as have property changed notification as well. I the use the following extension methods to remove those properties. I think I actually got the original code from the article that you linked.
public static class DbContextExtensions
{
public static void RemoveProperties(this ODataEntry entry, params string[] propertyNames)
{
var properties = entry.Properties as List<ODataProperty>;
if (properties == null)
{
properties = new List<ODataProperty>(entry.Properties);
}
var propertiesToRemove = properties.Where(p => propertyNames.Any(rp => rp == p.Name));
foreach (var propertyToRemove in propertiesToRemove.ToArray())
{
properties.Remove(propertyToRemove);
}
entry.Properties = properties;
}
public static DataServiceClientResponsePipelineConfiguration RemoveProperties<T>(this DataServiceClientResponsePipelineConfiguration responsePipeline, Func<string, Type> resolveType, params string[] propertiesToRemove)
{
return responsePipeline.OnEntryEnded((args) =>
{
Type resolvedType = resolveType(args.Entry.TypeName);
if (resolvedType != null && typeof(T).IsAssignableFrom(resolvedType))
{
args.Entry.RemoveProperties(propertiesToRemove);
}
});
}
public static DataServiceClientRequestPipelineConfiguration RemoveProperties<T>(this DataServiceClientRequestPipelineConfiguration requestPipeline, params string[] propertiesToRemove)
{
return requestPipeline.OnEntryStarting((args) =>
{
if (typeof(T).IsAssignableFrom(args.Entity.GetType()))
{
args.Entry.RemoveProperties(propertiesToRemove);
}
});
}
}
Notice that in the method below it is hooking OnEntryStarted. The code in the article hooks OnEntryEnded which worked for me at one point and then broke when I updated to a newer version of ODataClient. OnEntryStarted is the way to go in this method.
public static DataServiceClientRequestPipelineConfiguration RemoveProperties<T>(this DataServiceClientRequestPipelineConfiguration requestPipeline, params string[] propertiesToRemove)
{
return requestPipeline.OnEntryStarting((args) =>
{
if (typeof(T).IsAssignableFrom(args.Entity.GetType()))
{
args.Entry.RemoveProperties(propertiesToRemove);
}
});
}
I also created a partial class for the Container as well and implement the partial method OnContextCreated. This is where you use the extension methods to remove the properties that won't get sent to the server.
partial void OnContextCreated()
{
Configurations.RequestPipeline.RemoveProperties<Customer>(new string[] { "FullName", "VersionDetails" });
Configurations.RequestPipeline.RemoveProperties<SomeOtherType>(new string[] { "IsChecked", "IsReady" });
}
Make sure that your partial classes and the DBContextExtensions class are in the same namespace as our container and everything should just work.
Hope that helps.

Autofac and injection of instances

In a ASP.NET MVC project I'm working on I have the following piece of code that basically inject instances to specific methods within my assemblies.
So in the application root I have a class that register the instances like this and finally handles the injection.
ApplicationServiceProvider serviceProvider = ApplicationServiceProvider.CreateDefaultProvider();
serviceProvider.RegisterInstance(GlobalConfiguration.Configuration);
serviceProvider.RegisterInstance(GlobalFilters.Filters);
serviceProvider.RegisterInstance(RouteTable.Routes);
serviceProvider.RegisterInstance(BundleTable.Bundles);
serviceProvider.Distribute();
Now when I want to access these instances from the assemblies, I have to create some handler (method) and mark it with the following attribute 'ApplicationServiceHandler' like in the following example.
[ContractVerification(false)]
public static class RouteConfiguration
{
[ApplicationServiceHandler]
public static void Register(RouteCollection routes)
{
}
}
This is part of the extensibility layer in the project which is currently working pretty good.
Now, I'm new to Autofac and I wonder whether I can use Autofac to do the work for me rather than using my own implementation (which I provided below) that probably does it less efficient and handles less cases that are already covered by Autofac.
I noticed Autofac have a RegisterInstance method but I'm not sure how to tell it to inject the instances to methods flagged with 'ApplicationServiceHandler' attribute, I'm not not sure it's even the correct method but based on the name it seems like the right one.
Any kind of help is greatly appreciated, thank you.
EDIT: Here is the code that I'm using to achieve this without Autofac in my project.
ApplicationServiceHandlerAttribute.cs
[AttributeUsage(AttributeTargets.Method, AllowMultiple = false)]
public sealed class ApplicationServiceHandlerAttribute : Attribute
{
}
ApplicationServiceHandler.cs
public sealed class ApplicationServiceHandler
{
private readonly MethodInfo _method;
private readonly object[] _args;
public ApplicationServiceHandler(MethodInfo method, object[] args)
{
Contract.Requires(method != null);
Contract.Requires(args != null);
_method = method;
_args = args;
}
public void Invoke()
{
_method.Invoke(null, _args);
}
[ContractInvariantMethod]
[System.Diagnostics.CodeAnalysis.SuppressMessage("Microsoft.Performance", "CA1822:MarkMembersAsStatic", Justification = "Required for code contracts.")]
private void ObjectInvariant()
{
Contract.Invariant(_method != null);
Contract.Invariant(_args != null);
}
}
ApplicationServiceProvider.cs
public sealed class ApplicationServiceProvider
{
private readonly IEnumerable<Assembly> _assemblies;
private readonly Dictionary<Type, object> _instances;
public ApplicationServiceProvider(IEnumerable<Assembly> assemblies)
{
Contract.Requires(assemblies != null);
_assemblies = assemblies;
_instances = new Dictionary<Type, object>();
}
public static ApplicationServiceProvider CreateDefaultProvider()
{
Contract.Ensures(Contract.Result<ApplicationServiceProvider>() != null);
return new ApplicationServiceProvider(PackageLoader.ReferencedAssemblies);
}
public void Distribute()
{
foreach (var handler in GetHandlers())
{
Contract.Assume(handler != null);
handler.Invoke();
}
}
public IEnumerable<ApplicationServiceHandler> GetHandlers()
{
Contract.Ensures(Contract.Result<IEnumerable<ApplicationServiceHandler>>() != null);
if (_instances.Count == 0)
{
yield break;
}
foreach (var asm in _assemblies)
{
IEnumerable<MethodInfo> methods = GetMethods(asm);
foreach (var method in methods)
{
ParameterInfo[] #params = method.GetParameters();
if (#params.Length > 0)
{
int instanceCount = 0;
object[] args = new object[#params.Length];
for (int i = 0; i < #params.Length; i++)
{
ParameterInfo param = #params[i];
var instance = GetInstance(param);
if (instance != null)
{
instanceCount++;
args[i] = instance;
}
}
if (instanceCount > 0)
{
yield return new ApplicationServiceHandler(method, args);
}
}
}
}
}
public bool RegisterInstance(object instance)
{
Contract.Requires(instance != null);
return AddInstance(instance);
}
private static ApplicationServiceHandlerAttribute GetApplicationServiceHandlerAttribute(MethodInfo method)
{
ApplicationServiceHandlerAttribute attribute = null;
try
{
attribute = method.GetCustomAttribute<ApplicationServiceHandlerAttribute>(false);
}
catch (TypeLoadException)
{
// We don't need to do anything here for now.
}
return attribute;
}
private static IEnumerable<Type> GetDefinedTypes(Assembly assembly)
{
Contract.Requires(assembly != null);
Contract.Ensures(Contract.Result<IEnumerable<Type>>() != null);
try
{
return assembly.DefinedTypes;
}
catch (ReflectionTypeLoadException ex)
{
return ex.Types.Where(type => type != null);
}
}
/// <summary>
/// Gets the methods that are marked with <see cref="ApplicationServiceHandlerAttribute"/> from the assembly.
/// </summary>
/// <remarks>
/// Eyal Shilony, 21/11/2012.
/// </remarks>
/// <param name="assembly">
/// The assembly.
/// </param>
/// <returns>
/// The methods that are marked with <see cref="ApplicationServiceHandlerAttribute"/> from the assembly.
/// </returns>
private static IEnumerable<MethodInfo> GetMethods(Assembly assembly)
{
Contract.Requires(assembly != null);
Contract.Ensures(Contract.Result<IEnumerable<MethodInfo>>() != null);
const TypeAttributes STATIC_TYPE_ATTRIBUTES = TypeAttributes.AutoLayout | TypeAttributes.AnsiClass | TypeAttributes.Class | TypeAttributes.Public | TypeAttributes.Abstract | TypeAttributes.Sealed | TypeAttributes.BeforeFieldInit;
var methods = (from type in GetDefinedTypes(assembly)
where type.Attributes == STATIC_TYPE_ATTRIBUTES
from method in type.GetMethods().AsParallel()
where GetApplicationServiceHandlerAttribute(method) != null
select method).ToArray();
return methods;
}
private bool AddInstance(object instance)
{
Type type = instance.GetType();
return AddInstance(type, instance);
}
private bool AddInstance(Type type, object instance)
{
if (!_instances.ContainsKey(type))
{
_instances.Add(type, instance);
return true;
}
return false;
}
private object GetInstance(ParameterInfo param)
{
object instance = null;
Type paramType = param.ParameterType;
if (_instances.ContainsKey(paramType))
{
instance = _instances[paramType];
}
else
{
foreach (var type in _instances.Keys.Where(type => type.IsSubclassOf(paramType)))
{
instance = _instances[type];
break;
}
}
return instance;
}
}
i hope , i have understood you correctly.if what you mean is marking a class as dependency with attributes then you can do it by creating custom attribute.following is an example of implementing such an attribute :
public class DependencyAttribute : Attribute
{
public DependencyAttribute()
{
}
//The type of service the attributed class represents
public Type ServiceType { get; set; }
//Optional key to associate with the service
public string Key { get; set; }
public virtual void RegisterService(AttributeInfo<DependencyAttribute> attributeInfo, IContainer container)
{
Type serviceType = attributeInfo.Attribute.ServiceType ?? attributeInfo.DecoratedType;
Containerbuilder builder = new ContainerBuilder();
builder.RegisterType(attributeInfo.DecoratedType).As(serviceType).Keyed(
attributeInfo.Attribute.Key ?? attributeInfo.DecoratedType.FullName);
builder.Update(container)
}
}
then you must find all types marked with this attribute and call the RegisterService method of these attributes.
public class DependencyAttributeRegistrator
{
public DependencyAttributeRegistrator()
{
}
public IEnumerable<AttributeInfo<DependencyAttribute>> FindServices()
{
//replace this line with you'r own
var types = Assembly.GetExecutingAssembly().GetTypes();
foreach (Type type in types)
{
var attributes = type.GetCustomAttributes(typeof(DependencyAttribute), false);
foreach (DependencyAttribute attribute in attributes)
{
yield return new AttributeInfo<DependencyAttribute> { Attribute = attribute, DecoratedType = type };
}
}
}
public void RegisterServices(IEnumerable<AttributeInfo<DependencyAttribute>> services)
{
foreach (var info in services)
{
//replace following two line with you'r own global container
var builder = new ContainerBuilder();
IContainer container = builder.Build();
info.Attribute.RegisterService(info, container);
}
}
}

MVC3 Controller with specific JsonConverter

Here's the setup:
I have some MVC Controllers that are intended to be consumed by jQuery ajax requests. A normal request would seem somewhat like this:
$.ajax("/Solicitor/AddSolicitorToApplication", {
data: putData,
type: "POST", contentType: "application/json",
success: function (result) {
//My success callback
}
}
});
My controller looks like this:
[HttpPost]
public ActionResult InsertLoanApplication(MortgageLoanApplicationViewModel vm)
{
var mortgageLoanDTO = vm.MapToDTO();
return Json(_mortgageLoanService.UpdateMortgageLoanApplication(mortgageLoanDTO), JsonRequestBehavior.DenyGet);
}
This works perfectly fine with most objects passed to the controller, except that in this specific case one of the properties of the object being passed needs to be deserialized in a specific way.
I've added a JsonConverter that I've used previously with the MVC4 Web API, but in this case I need to apply it to regular mvc controllers.
I tried registering the JsonConverter in my global.asax like this:
GlobalConfiguration.Configuration.Formatters.JsonFormatter.SerializerSettings.Converters.Add(new GrizlyStringConverter());
But so far haven't been able to deserialize the object.
You should replace the built-in JsonValueProviderFactory class with a custom one if you want to use Json.NET when binding JSON requests to view models.
You could write one as shown in this gist:
public sealed class JsonDotNetValueProviderFactory : ValueProviderFactory
{
public override IValueProvider GetValueProvider(ControllerContext controllerContext)
{
if (controllerContext == null)
{
throw new ArgumentNullException("controllerContext");
}
if (!controllerContext.HttpContext.Request.ContentType.StartsWith("application/json", StringComparison.OrdinalIgnoreCase))
{
return null;
}
using (var reader = new StreamReader(controllerContext.HttpContext.Request.InputStream))
{
var bodyText = reader.ReadToEnd();
return String.IsNullOrEmpty(bodyText)
? null :
new DictionaryValueProvider<object>(
JsonConvert.DeserializeObject<ExpandoObject>(
bodyText,
new ExpandoObjectConverter()
),
CultureInfo.CurrentCulture
);
}
}
}
and then replace the built-in with your custom one in Application_Start:
ValueProviderFactories.Factories.Remove(
ValueProviderFactories
.Factories
.OfType<JsonValueProviderFactory>()
.FirstOrDefault()
);
ValueProviderFactories.Factories.Add(new JsonDotNetValueProviderFactory());
That's it. Now you are using Json.Net instead of the JavaScriptSerializer for the incoming JSON requests.
The modified version:
using System;
using System.Collections;
using System.Collections.Generic;
using System.Dynamic;
using System.Globalization;
using System.IO;
using System.Web.Mvc;
using Newtonsoft.Json;
using Newtonsoft.Json.Converters;
namespace MvcJsonNetTests.Utils
{
public class JsonNetValueProviderFactory : ValueProviderFactory
{
public JsonNetValueProviderFactory()
{
Settings = new JsonSerializerSettings
{
ReferenceLoopHandling = ReferenceLoopHandling.Error,
Converters = { new ExpandoObjectConverter() }
};
}
public JsonSerializerSettings Settings { get; set; }
public override IValueProvider GetValueProvider(ControllerContext controllerContext)
{
if (controllerContext == null)
throw new ArgumentNullException("controllerContext");
if (controllerContext.HttpContext == null ||
controllerContext.HttpContext.Request == null ||
controllerContext.HttpContext.Request.ContentType == null)
{
return null;
}
if (!controllerContext.HttpContext.Request.ContentType.StartsWith(
"application/json", StringComparison.OrdinalIgnoreCase))
{
return null;
}
if (!controllerContext.HttpContext.Request.IsAjaxRequest())
{
return null;
}
using (var streamReader = new StreamReader(controllerContext.HttpContext.Request.InputStream))
{
using (var jsonReader = new JsonTextReader(streamReader))
{
if (!jsonReader.Read())
return null;
var jsonSerializer = JsonSerializer.Create(this.Settings);
Object jsonObject;
switch (jsonReader.TokenType)
{
case JsonToken.StartArray:
jsonObject = jsonSerializer.Deserialize<List<ExpandoObject>>(jsonReader);
break;
default:
jsonObject = jsonSerializer.Deserialize<ExpandoObject>(jsonReader);
break;
}
var backingStore = new Dictionary<string, object>(StringComparer.OrdinalIgnoreCase);
addToBackingStore(backingStore, String.Empty, jsonObject);
return new DictionaryValueProvider<object>(backingStore, CultureInfo.CurrentCulture);
}
}
}
private static void addToBackingStore(IDictionary<string, object> backingStore, string prefix, object value)
{
var dictionary = value as IDictionary<string, object>;
if (dictionary != null)
{
foreach (var entry in dictionary)
{
addToBackingStore(backingStore, makePropertyKey(prefix, entry.Key), entry.Value);
}
return;
}
var list = value as IList;
if (list != null)
{
for (var index = 0; index < list.Count; index++)
{
addToBackingStore(backingStore, makeArrayKey(prefix, index), list[index]);
}
return;
}
backingStore[prefix] = value;
}
private static string makeArrayKey(string prefix, int index)
{
return prefix + "[" + index.ToString(CultureInfo.InvariantCulture) + "]";
}
private static string makePropertyKey(string prefix, string propertyName)
{
return (string.IsNullOrWhiteSpace(prefix)) ? propertyName : prefix + "." + propertyName;
}
}
}
Also to register it at the right index:
public static void RegisterFactory()
{
var defaultJsonFactory = ValueProviderFactories.Factories
.OfType<JsonValueProviderFactory>().FirstOrDefault();
var index = ValueProviderFactories.Factories.IndexOf(defaultJsonFactory);
ValueProviderFactories.Factories.Remove(defaultJsonFactory);
ValueProviderFactories.Factories.Insert(index, new JsonNetValueProviderFactory());
}

Serialize IList property on model when passed into Html.ActionLink

I'm trying to generate an Html.ActionLink with the following viewmodel:
public class SearchModel
{
public string KeyWords {get;set;}
public IList<string> Categories {get;set;}
}
To generate my link I use the following call:
#Html.ActionLink("Index", "Search", Model)
Where Model is an instance of the SearchModel
The link generated is something like this:
http://www.test.com/search/index?keywords=bla&categories=System.Collections.Generic.List
Because it obviously is only calling the ToString method on every property.
What I would like to see generate is this:
http://www.test.com/search/index?keywords=bla&categories=Cat1&categories=Cat2
Is there any way I can achieve this by using Html.ActionLink
In MVC 3 you're just out of luck because the route values are stored in a RouteValueDictionary that as the name implies uses a Dictionary internally which makes it not possible to have multiple values associated to a single key. The route values should probably be stored in a NameValueCollection to support the same behavior as the query string.
However, if you can impose some constraints on the categories names and you're able to support a query string in the format:
http://www.test.com/search/index?keywords=bla&categories=Cat1|Cat2
then you could theoretically plug it into Html.ActionLink since MVC uses TypeDescriptor which in turn is extensible at runtime. The following code is presented to demonstrate it's possible, but I would not recommend it to be used, at least without further refactoring.
Having said that, you would need to start by associating a custom type description provider:
[TypeDescriptionProvider(typeof(SearchModelTypeDescriptionProvider))]
public class SearchModel
{
public string KeyWords { get; set; }
public IList<string> Categories { get; set; }
}
The implementation for the provider and the custom descriptor that overrides the property descriptor for the Categories property:
class SearchModelTypeDescriptionProvider : TypeDescriptionProvider
{
public override ICustomTypeDescriptor GetTypeDescriptor(
Type objectType, object instance)
{
var searchModel = instance as SearchModel;
if (searchModel != null)
{
var properties = new List<PropertyDescriptor>();
properties.Add(TypeDescriptor.CreateProperty(
objectType, "KeyWords", typeof(string)));
properties.Add(new ListPropertyDescriptor("Categories"));
return new SearchModelTypeDescriptor(properties.ToArray());
}
return base.GetTypeDescriptor(objectType, instance);
}
}
class SearchModelTypeDescriptor : CustomTypeDescriptor
{
public SearchModelTypeDescriptor(PropertyDescriptor[] properties)
{
this.Properties = properties;
}
public PropertyDescriptor[] Properties { get; set; }
public override PropertyDescriptorCollection GetProperties()
{
return new PropertyDescriptorCollection(this.Properties);
}
}
Then we would need the custom property descriptor to be able to return a custom value in GetValue which is called internally by MVC:
class ListPropertyDescriptor : PropertyDescriptor
{
public ListPropertyDescriptor(string name)
: base(name, new Attribute[] { }) { }
public override bool CanResetValue(object component)
{
return false;
}
public override Type ComponentType
{
get { throw new NotImplementedException(); }
}
public override object GetValue(object component)
{
var property = component.GetType().GetProperty(this.Name);
var list = (IList<string>)property.GetValue(component, null);
return string.Join("|", list);
}
public override bool IsReadOnly { get { return false; } }
public override Type PropertyType
{
get { throw new NotImplementedException(); }
}
public override void ResetValue(object component) { }
public override void SetValue(object component, object value) { }
public override bool ShouldSerializeValue(object component)
{
throw new NotImplementedException();
}
}
And finally to prove that it works a sample application that mimics the MVC route values creation:
static void Main(string[] args)
{
var model = new SearchModel { KeyWords = "overengineering" };
model.Categories = new List<string> { "1", "2", "3" };
var properties = TypeDescriptor.GetProperties(model);
var dictionary = new Dictionary<string, object>();
foreach (PropertyDescriptor p in properties)
{
dictionary.Add(p.Name, p.GetValue(model));
}
// Prints: KeyWords, Categories
Console.WriteLine(string.Join(", ", dictionary.Keys));
// Prints: overengineering, 1|2|3
Console.WriteLine(string.Join(", ", dictionary.Values));
}
Damn, this is probably the longest answer I ever give here at SO.
with linq of course...
string.Join("", Model.Categories.Select(c=>"&categories="+c))

attribute dependent on another field

In a model of my ASP.NET MVC application I would like validate a textbox as required only if a specific checkbox is checked.
Something like
public bool retired {get, set};
[RequiredIf("retired",true)]
public string retirementAge {get, set};
How can I do that?
Thank you.
Take a look at this: http://blogs.msdn.com/b/simonince/archive/2010/06/04/conditional-validation-in-mvc.aspx
I've modded the code somewhat to suit my needs. Perhaps you benefit from those changes as well.
public class RequiredIfAttribute : ValidationAttribute
{
private RequiredAttribute innerAttribute = new RequiredAttribute();
public string DependentUpon { get; set; }
public object Value { get; set; }
public RequiredIfAttribute(string dependentUpon, object value)
{
this.DependentUpon = dependentUpon;
this.Value = value;
}
public RequiredIfAttribute(string dependentUpon)
{
this.DependentUpon = dependentUpon;
this.Value = null;
}
public override bool IsValid(object value)
{
return innerAttribute.IsValid(value);
}
}
public class RequiredIfValidator : DataAnnotationsModelValidator<RequiredIfAttribute>
{
public RequiredIfValidator(ModelMetadata metadata, ControllerContext context, RequiredIfAttribute attribute)
: base(metadata, context, attribute)
{ }
public override IEnumerable<ModelClientValidationRule> GetClientValidationRules()
{
// no client validation - I might well blog about this soon!
return base.GetClientValidationRules();
}
public override IEnumerable<ModelValidationResult> Validate(object container)
{
// get a reference to the property this validation depends upon
var field = Metadata.ContainerType.GetProperty(Attribute.DependentUpon);
if (field != null)
{
// get the value of the dependent property
var value = field.GetValue(container, null);
// compare the value against the target value
if ((value != null && Attribute.Value == null) || (value != null && value.Equals(Attribute.Value)))
{
// match => means we should try validating this field
if (!Attribute.IsValid(Metadata.Model))
// validation failed - return an error
yield return new ModelValidationResult { Message = ErrorMessage };
}
}
}
}
Then use it:
public DateTime? DeptDateTime { get; set; }
[RequiredIf("DeptDateTime")]
public string DeptAirline { get; set; }
Just use the Foolproof validation library that is available on Codeplex:
https://foolproof.codeplex.com/
It supports, amongst others, the following "requiredif" validation attributes / decorations:
[RequiredIf]
[RequiredIfNot]
[RequiredIfTrue]
[RequiredIfFalse]
[RequiredIfEmpty]
[RequiredIfNotEmpty]
[RequiredIfRegExMatch]
[RequiredIfNotRegExMatch]
To get started is easy:
Download the package from the provided link
Add a reference to the included .dll file
Import the included javascript files
Ensure that your views references the included javascript files from within its HTML for unobtrusive javascript and jquery validation.
Using NuGet Package Manager I intstalled this: https://github.com/jwaliszko/ExpressiveAnnotations
And this is my Model:
using ExpressiveAnnotations.Attributes;
public bool HasReferenceToNotIncludedFile { get; set; }
[RequiredIf("HasReferenceToNotIncludedFile == true", ErrorMessage = "RelevantAuditOpinionNumbers are required.")]
public string RelevantAuditOpinionNumbers { get; set; }
I guarantee you this will work!
I have not seen anything out of the box that would allow you to do this.
I've created a class for you to use, it's a bit rough and definitely not flexible.. but I think it may solve your current problem. Or at least put you on the right track.
using System;
using System.Collections.Generic;
using System.Linq;
using System.Web;
using System.ComponentModel.DataAnnotations;
using System.Globalization;
namespace System.ComponentModel.DataAnnotations
{
[AttributeUsage(AttributeTargets.Class, AllowMultiple = true, Inherited = true)]
public sealed class RequiredIfAttribute : ValidationAttribute
{
private const string _defaultErrorMessage = "'{0}' is required";
private readonly object _typeId = new object();
private string _requiredProperty;
private string _targetProperty;
private bool _targetPropertyCondition;
public RequiredIfAttribute(string requiredProperty, string targetProperty, bool targetPropertyCondition)
: base(_defaultErrorMessage)
{
this._requiredProperty = requiredProperty;
this._targetProperty = targetProperty;
this._targetPropertyCondition = targetPropertyCondition;
}
public override object TypeId
{
get
{
return _typeId;
}
}
public override string FormatErrorMessage(string name)
{
return String.Format(CultureInfo.CurrentUICulture, ErrorMessageString, _requiredProperty, _targetProperty, _targetPropertyCondition);
}
public override bool IsValid(object value)
{
bool result = false;
bool propertyRequired = false; // Flag to check if the required property is required.
PropertyDescriptorCollection properties = TypeDescriptor.GetProperties(value);
string requiredPropertyValue = (string) properties.Find(_requiredProperty, true).GetValue(value);
bool targetPropertyValue = (bool) properties.Find(_targetProperty, true).GetValue(value);
if (targetPropertyValue == _targetPropertyCondition)
{
propertyRequired = true;
}
if (propertyRequired)
{
//check the required property value is not null
if (requiredPropertyValue != null)
{
result = true;
}
}
else
{
//property is not required
result = true;
}
return result;
}
}
}
Above your Model class, you should just need to add:
[RequiredIf("retirementAge", "retired", true)]
public class MyModel
In your View
<%= Html.ValidationSummary() %>
Should show the error message whenever the retired property is true and the required property is empty.
Hope this helps.
Try my custom validation attribute:
[ConditionalRequired("retired==true")]
public string retirementAge {get, set};
It supports multiple conditions.

Resources