Validation using MediatR's Pipeline Behaviors and FluentValidation | Clean .NET Core 

DTO validation makes sense to me, considering validating domain models means your API would be returning bad requests for a different model than what was sent to your API. In the video here, it's simple because there are just two properties in the request models, but if your DTOs don't match the domain models, then your bad requests will be confusing for your API callers. In theory, it might make sense to have two validation logics (DTO validation + domain model validation), but in practice, that sounds really overkill unless you've got a very complex solution.

I would like to know as well, I know there is an option to validate manually via a Result request object, but it would be nice to know if there is a method with less boiler plate and same performance. The only way I can think of is the Asp.Net filters method, but that is not related to mediatr.

you all probably dont care at all but does someone know a tool to get back into an Instagram account..? I stupidly forgot my login password. I appreciate any tips you can give me.

@Jason Leandro instablaster ;)

@Abdullah Danny thanks so much for your reply. I found the site through google and im waiting for the hacking stuff now. Looks like it's gonna take a while so I will get back to you later with my results.

@Abdullah Danny It worked and I finally got access to my account again. Im so happy:D Thanks so much you saved my account !

@Jason Leandro Happy to help =)

i returned a solid state result contains my result info instead of throwing exceptions Result : is my Return Result Class contains a method named Feiler to hold the error information this code is not optimized yet but i hope it the startup answer for not using exceptions if (failures.Any()) { var errors = failures.Select(p => p.ErrorMessage).ToArray(); if(typeof(TResponse).Name.Contains("Result")) { MethodInfo builderMethod = typeof(TResponse).GetMethod("Failure", BindingFlags.Static | BindingFlags.Public); var res = (TResponse)builderMethod.Invoke(null, new object[] { errors }); return res; } }

@Elfocrash please answer thsi question.

I need an answer on this question as well!

I think I mention that in the video but that was a while ago so I could be wrong. Middleware will validate your API level model. You shouldn’t really use them for domain related concerns. Middleware should catch things related to the api request for example “this field is missing from the body or this should be a string”. Pipelines sea with the domain model and validate domain related concerns. In reality you need both.

@@nickchapsas Thanks for the videos they are awesome, I am little confused on this video, Should domain level cross cutting concerns reside in domain layer or at application layer ?Domain layer exceptions (not errors) can be handled by giving a proper messge to consumer (Bad Request), however for application level or persistance level errors i want to handle in a different way(let;s say common error message and logging ), Should'nt a business level dependency be a part of the domain layer ?

Why would you use middlewares? You would end up adding loads of Middlewares for all kinds of domain-centric scenarios for every single request. So, for any one request, almost all of those middlewares would not be relevant. This is not a good approach if you care about performance. (This is not a good approach anyway 😁)

It's probably MS Whiteboard :)

Depends on your performance requirements

@@nickchapsas Can you make a course/ video on Validation, especially focussing on which responsibility lies in which levels of validation (Domain, Api)? I checked your video on CleanArchitecture and ValueObjects but I am a bit befuddled by that topic now. Thank you for reading :-)

He's using Jetbrains' Rider IDE

It doesn't. The synchronous way has actually been completely removed for a long time now.

@@nickchapsas I didn't know that. Thanks. Is it possible to setup multiple handler for a single command? And is it possible to setup multiple handler for single iNotification? Will both handler be executed same time?

@@dfytq Yes you can. Notifications is a one to many relationships and all notification handlers will run when a notification is sent

Rider from JetBrains

@ , thanks for your reply , just found a link for the comparison www.jetbrains.com/rider/compare/rider-vs-visual-studio/

The advantage is that there is a clear separation between your API layer and your domain layer. I you pass the same object then your API and your domain are coupled and you are risking making breaking changes on the API level.

@@nickchapsas In my project i pass to the API a "CreateValueCommand" that is different from my domain model "Value". Are you using the CreateCustomerOrderCommand as a Domain model in this video?

That’s correct

@@nickchapsas appreciated, thank you

Hello Somnath. The code is available to my Patreons

I would wrapp all 3 of those things in a UpdateCustomerOrderRequest, which contains all those properties with their attributes and create an AbstractValidator. Then you can simply validate on that.

@@nickchapsas Does it look like this? End up with multiple classes (the Command and the wrapped class) inheriting from IRequest Bit confused about the use of a Command object in this scenario where the Command represents the body but the domain logic also needs the Route/Query params Should the UpdateCustomerOrderRequest become the Command object? (is this the best place to ask code questions?) public class UpdateCustomerOrderRequest: IRequest //should this be the Command object? { [FromRoute] internal Guid OrderId { get; set; } [FromBody] internal UpdateCustomerOrderCommand Command { get; set; } //should this be a POCO? } public class UpdateCustomerOrderCommand : IRequest { public Guid CustomerId { get; set; } public Guid ProductId { get; set; } } [HttpPut("{orderId}")] public async Task UpdateOrder(UpdateCustomerOrderRequest request) { await _mediator.Send(request); return Ok(); }

I would use unit tests and integration tests

chg it to dotnetcore 3.0

I use services.AddValidatorsFromAssembly(Assembly.Load("full.assembly.name"));

The source code is available by becoming a Patreon member

The code for all my videos is available to my Patreons

public class ValidationFilter : IAsyncActionFilter { public async Task OnActionExecutionAsync(ActionExecutingContext context, ActionExecutionDelegate next) { if (!context.ModelState.IsValid) { var modelstateErrors = context.ModelState.Where(x => x.Value.Errors.Count > 0) .ToDictionary(k => k.Key, v => v.Value.Errors.Select(x => x.ErrorMessage).ToArray()); var validationResponse = new ValidationResponse(); validationResponse.GlobalValidationMessage = "Invalid form value"; foreach (var error in modelstateErrors) { foreach (var message in error.Value) { validationResponse.ValidationErrors.Add(new ValidationErrorModel { FieldName = error.Key, Message = message }); } } context.Result = new BadRequestObjectResult(validationResponse); return; } await next(); } }

Please send me your GitHub username and I will manually invite you

@@nickchapsas I spoke too soon. I was able to access it! Thank you!

MediatR makes it very easy to inject fakes during unit testing.