Strange behaviour on IEqualityComparer

Question

I'm trying to implement an IEqualityComparer for my object that basically detects if an object is older that another one. The following simpler example will synthesises what i'm trying to accomplish:

class Program
{
    static void Main(string[] args)
    {
        var authorsList = new List<Author>()
        {
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=11 },
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=20 },
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=12 },
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=14 },
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=12 },
            new Author{ Firstname = "Fred", Lastname = "Smith", Age=12 },
            new Author{ Firstname = "Trevor", Lastname = "Smith", Age=15 },
            new Author{ Firstname = "Brian", Lastname = "Smith", Age=11 },
            new Author{ Firstname = "Billy", Lastname = "Smith", Age=11 },
        };
        var authorsListExcept = new List<Author>()
        {
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=12 },
            new Author{ Firstname = "Fred", Lastname = "Smith", Age=12 },
        };
        var authorsList2 = authorsList
            .Except(authorsListExcept, new AuthorUpdatedComparer()).ToList();
    }
}

class Author
{
    public string Firstname { get; set; }
    public string Lastname { get; set; }
    public int Age { get; set; }
}

class AuthorUpdatedComparer : IEqualityComparer<Author>
{
    public bool Equals(Author x, Author y)
    {
        return x.Age >= y.Age;
    }

    public int GetHashCode(Author obj)
    {
        //Check whether the object is null 
        if (Object.ReferenceEquals(obj, null)) return 0;

        int FirstnameHash = (obj.Firstname ?? "").GetHashCode();
        int LastnameHash = (obj.Lastname ?? "").GetHashCode();
        int finalResult = FirstnameHash ^ LastnameHash;

        return finalResult;
    }
}

My authorsList2 result would be: * Bob Smith with age 20 * Bom Smith with age 14 * Trevor Smith with age 15 * Brian Smith with age 11 * Billy Smith with age 11

But instead of this the Bob Smith with age 14 is not included. When debugging I reached the conclusion that the Comparer after included Bob Smith with Age 20 start to using it has a comparer excluding then after all the Bob's younger than 20.

This is a strange behavior in my point of view it should only excluded the ones that are younger or with the same age to those included on the authorsListExcept. I tried to read msdn documentation and what I want it should be supposed to happen: font: http://msdn.microsoft.com/en-us/library/bb336390(v=vs.100).aspx

Anyone can help me? Thanks, Hugo Salgado

Answer 1

The following LINQ query provides the result you expect:

var result = 
    authorsList.GroupBy(x => Tuple.Create(x.Firstname, x.Lastname))
               .SelectMany(g => g.Where(x => authorsListExcept.All(e => e.Firstname != x.Firstname || e.Lastname != x.Lastname || e.Age < x.Age)));

The following query also produces this result. It should perform better:

var result = 
    authorsList.GroupBy(x => Tuple.Create(x.Firstname, x.Lastname))
               .GroupJoin(authorsListExcept, x => x.Key,
                          x => Tuple.Create(x.Firstname, x.Lastname),
                          (a, e) => a.Where(x => x.Age > e.Select(z => z.Age)
                                                          .DefaultIfEmpty(0)
                                                          .Max()))
               .SelectMany(x => x)

And a third option (the same as the previous but in query syntax):

var result = 
    (from a in authorsList
    group a by Tuple.Create(a.Firstname, a.Lastname) into g
    join e in authorsListExcept on g.Key equals Tuple.Create(e.Firstname, e.Lastname) into er
    from age in er.Select(x => x.Age).DefaultIfEmpty()
    select g.Where(x => x.Age > age)).SelectMany(x => x);

---

The interface IEqualityComparer<T> is there to check for equality. It has nothing to do with any ordering. As such, you can't use it in the way you try.

In general: An implementation of this interface should always use the exact same set of properties in both the GetHashCode implementation and the Equals method.

Answer 2

If i have understood what you want try this.

class Program
{
    static void Main(string[] args)
    {
        var authorsList = new List<Author>()
        {
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=11 },
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=20 },
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=12 },
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=14 },
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=12 },
            new Author{ Firstname = "Fred", Lastname = "Smith", Age=12 },
            new Author{ Firstname = "Trevor", Lastname = "Smith", Age=15 },
            new Author{ Firstname = "Brian", Lastname = "Smith", Age=11 },
            new Author{ Firstname = "Billy", Lastname = "Smith", Age=11 },
        };
        var authorsListExcept = new List<Author>()
        {
            new Author{ Firstname = "Bob", Lastname = "Smith", Age=12 },
            new Author{ Firstname = "Fred", Lastname = "Smith", Age=12 },
        };

        var authorsList2 = authorsList.Where(x => !authorsListExcept.Any(y => y.Firstname == x.Firstname && y.Lastname == x.Lastname && x.Age <= y.Age));
    }
}

public class Author
{
    public string Firstname { get; set; }
    public string Lastname { get; set; }
    public int Age { get; set; }
}