Generic result class with C#

Generic result class with C#

Summary

In this post, I discuss a way to create a result class that can be used for method calls across multiple layers in your application. I discuss the need for such an approach and my take on the solution. I welcome recommendations to make this solution better.

Introduction

It has become the norm that when one starts creating a .net solution, one follows a layered approach. Following the standard n-layered architecture, one creates a number of class libraries to represent the layers. Often times, I have come across the problem of getting the results from the lowest layer to the topmost layer. So, for example, I write a method that gets data from the database and populate an object and I need to return this object to the UI, the norm is to create a method signature that returns the object type. An issue arise when errors happen, how do I get to send these errors to the upper layers?

The following sections look at a solution I came to like and have been using in most of my solutions. Let’s assume I am writing an application to read manage student details. I will define some methods to manage this data and describe a way to use a generic result class I can use in all my methods. I will also describe a caveat to this solution and a work around.

IStudentRepository

public interface IStudentRepository

{

long AddStudent(Student newStudent);

Student GetStudent(long studentId);

bool UpdateStudent(Student updatedStudent);

}

Note that all my methods return a different data type. How should I handle errors that occur when trying to access my persistence layer? The solution I use id to create a generic result class that has the capability to carry errors as well as my results and any other piece of information. I also use a couple of other classes to help manage this.

Generic result

[DataContract]

    public class Result

    {

        public Result()

        {

            IsValid = true;

            Errors = new List<Error>();

        }

        [DataMember]

        public T Value { get; set; }

        [DataMember]

        public bool IsValid { get; set; }

        [DataMember]

        public List<Error> Errors { get; set; }

        [DataMember]

        public List<string> InfoMessages { get; set; }

    }

[DataContract]

    public class Error

    {

        [DataMember]

        public string ErrorCode { get; set; }

        [DataMember]

        public string ErrorMessage { get; set; }

        [DataMember]

        public ErrorType Severity { get; set; }

    }

The DataContract attibutes have been added allow this to be serialized in WCF for example. Now using this class, here is how I define the same interface above

public interface IStudentRepository

{

Result<long> AddStudent(Student newStudent);

Result<Student> GetStudent(long studentId);

Result<bool> UpdateStudent(Student updatedStudent);

}

Using this approach, in my code I first check if the result is valid, if it is, I can access the strongly typed Value property to get the results and pontentially any InfoMessages I can use for notifications. If the result is invalid, I can access the list of errors to get more information on what went wrong.

Caveat L

With this solution, you will realize that it will not work when your result type is a string, a work around is to wrap your string value in an object

I hope you find this useful in your solutions too.