Can you create a simple 'EqualityComparer' using a lambda expression

Question

Short question:

Is there a simple way in LINQ to objects to get a distinct list of objects from a list based on a key property on the objects.

Long question:

I am trying to do a Distinct() operation on a list of objects that have a key as one of their properties.

class GalleryImage {
   public int Key { get;set; }
   public string Caption { get;set; }
   public string Filename { get; set; }
   public string[] Tags {g et; set; }
}

I have a list of Gallery objects that contain GalleryImage[].

Because of the way the webservice works [sic] I have duplicates of the GalleryImage object. i thought it would be a simple matter to use Distinct() to get a distinct list.

This is the LINQ query I want to use :

var allImages = Galleries.SelectMany(x => x.Images);
var distinctImages = allImages.Distinct<GalleryImage>(new 
                     EqualityComparer<GalleryImage>((a, b) => a.id == b.id));

The problem is that EqualityComparer is an abstract class.

I dont want to :

• implement IEquatable on GalleryImage because it is generated
• have to write a separate class to implement IEqualityComparer as shown here

Is there a concrete implementation of EqualityComparer somewhere that I'm missing?

I would have thought there would be an easy way to get 'distinct' objects from a set based on a key.

Answer 1

(There are two solutions here - see the end for the second one):

My MiscUtil library has a ProjectionEqualityComparer class (and two supporting classes to make use of type inference).

Here's an example of using it:

EqualityComparer<GalleryImage> comparer = 
    ProjectionEqualityComparer<GalleryImage>.Create(x => x.id);

Here's the code (comments removed)

// Helper class for construction
public static class ProjectionEqualityComparer
{
    public static ProjectionEqualityComparer<TSource, TKey>
        Create<TSource, TKey>(Func<TSource, TKey> projection)
    {
        return new ProjectionEqualityComparer<TSource, TKey>(projection);
    }

    public static ProjectionEqualityComparer<TSource, TKey>
        Create<TSource, TKey> (TSource ignored,
                               Func<TSource, TKey> projection)
    {
        return new ProjectionEqualityComparer<TSource, TKey>(projection);
    }
}

public static class ProjectionEqualityComparer<TSource>
{
    public static ProjectionEqualityComparer<TSource, TKey>
        Create<TKey>(Func<TSource, TKey> projection)
    {
        return new ProjectionEqualityComparer<TSource, TKey>(projection);
    }
}

public class ProjectionEqualityComparer<TSource, TKey>
    : IEqualityComparer<TSource>
{
    readonly Func<TSource, TKey> projection;
    readonly IEqualityComparer<TKey> comparer;

    public ProjectionEqualityComparer(Func<TSource, TKey> projection)
        : this(projection, null)
    {
    }

    public ProjectionEqualityComparer(
        Func<TSource, TKey> projection,
        IEqualityComparer<TKey> comparer)
    {
        projection.ThrowIfNull("projection");
        this.comparer = comparer ?? EqualityComparer<TKey>.Default;
        this.projection = projection;
    }

    public bool Equals(TSource x, TSource y)
    {
        if (x == null && y == null)
        {
            return true;
        }
        if (x == null || y == null)
        {
            return false;
        }
        return comparer.Equals(projection(x), projection(y));
    }

    public int GetHashCode(TSource obj)
    {
        if (obj == null)
        {
            throw new ArgumentNullException("obj");
        }
        return comparer.GetHashCode(projection(obj));
    }
}

Second solution

To do this just for Distinct, you can use the DistinctBy extension in MoreLINQ:

    public static IEnumerable<TSource> DistinctBy<TSource, TKey>
        (this IEnumerable<TSource> source,
         Func<TSource, TKey> keySelector)
    {
        return source.DistinctBy(keySelector, null);
    }

    public static IEnumerable<TSource> DistinctBy<TSource, TKey>
        (this IEnumerable<TSource> source,
         Func<TSource, TKey> keySelector,
         IEqualityComparer<TKey> comparer)
    {
        source.ThrowIfNull("source");
        keySelector.ThrowIfNull("keySelector");
        return DistinctByImpl(source, keySelector, comparer);
    }

    private static IEnumerable<TSource> DistinctByImpl<TSource, TKey>
        (IEnumerable<TSource> source,
         Func<TSource, TKey> keySelector,
         IEqualityComparer<TKey> comparer)
    {
        HashSet<TKey> knownKeys = new HashSet<TKey>(comparer);
        foreach (TSource element in source)
        {
            if (knownKeys.Add(keySelector(element)))
            {
                yield return element;
            }
        }
    }

In both cases, ThrowIfNull looks like this:

public static void ThrowIfNull<T>(this T data, string name) where T : class
{
    if (data == null)
    {
        throw new ArgumentNullException(name);
    }
}

Answer 2

Building on Charlie Flowers' answer, you can create your own extension method to do what you want which internally uses grouping:

    public static IEnumerable<T> Distinct<T, U>(
        this IEnumerable<T> seq, Func<T, U> getKey)
    {
        return
            from item in seq
            group item by getKey(item) into gp
            select gp.First();
    }

You could also create a generic class deriving from EqualityComparer, but it sounds like you'd like to avoid this:

    public class KeyEqualityComparer<T,U> : IEqualityComparer<T>
    {
        private Func<T,U> GetKey { get; set; }

        public KeyEqualityComparer(Func<T,U> getKey) {
            GetKey = getKey;
        }

        public bool Equals(T x, T y)
        {
            return GetKey(x).Equals(GetKey(y));
        }

        public int GetHashCode(T obj)
        {
            return GetKey(obj).GetHashCode();
        }
    }

Answer 3

This is the best i can come up with for the problem in hand. Still curious whether theres a nice way to create a EqualityComparer on the fly though.

Galleries.SelectMany(x => x.Images).ToLookup(x => x.id).Select(x => x.First());

Create lookup table and take 'top' from each one

Note: this is the same as @charlie suggested but using ILookup - which i think is what a group must be anyway.

Answer 4

This idea is being debated here, and while I'm hoping the .NET Core team adopt a method to generate IEqualityComparer<T>s from lambda, I'd suggest you to please vote and comment on that idea, and use the following:

Usage:

IEqualityComparer<Contact> comp1 = EqualityComparerImpl<Contact>.Create(c => c.Name);
var comp2 = EqualityComparerImpl<Contact>.Create(c => c.Name, c => c.Age);

class Contact { public Name { get; set; } public Age { get; set; } }

Code:

public class EqualityComparerImpl<T> : IEqualityComparer<T>
{
  public static EqualityComparerImpl<T> Create(
    params Expression<Func<T, object>>[] properties) =>
    new EqualityComparerImpl<T>(properties);

  PropertyInfo[] _properties;
  EqualityComparerImpl(Expression<Func<T, object>>[] properties)
  {
    if (properties == null)
      throw new ArgumentNullException(nameof(properties));

    if (properties.Length == 0)
      throw new ArgumentOutOfRangeException(nameof(properties));

    var length = properties.Length;
    var extractions = new PropertyInfo[length];
    for (int i = 0; i < length; i++)
    {
      var property = properties[i];
      extractions[i] = ExtractProperty(property);
    }
    _properties = extractions;
  }

  public bool Equals(T x, T y)
  {
    if (ReferenceEquals(x, y))
      //covers both are null
      return true;
    if (x == null || y == null)
      return false;
    var len = _properties.Length;
    for (int i = 0; i < _properties.Length; i++)
    {
      var property = _properties[i];
      if (!Equals(property.GetValue(x), property.GetValue(y)))
        return false;
    }
    return true;
  }

  public int GetHashCode(T obj)
  {
    if (obj == null)
      return 0;

    var hashes = _properties
        .Select(pi => pi.GetValue(obj)?.GetHashCode() ?? 0).ToArray();
    return Combine(hashes);
  }

  static int Combine(int[] hashes)
  {
    int result = 0;
    foreach (var hash in hashes)
    {
      uint rol5 = ((uint)result << 5) | ((uint)result >> 27);
      result = ((int)rol5 + result) ^ hash;
    }
    return result;
  }

  static PropertyInfo ExtractProperty(Expression<Func<T, object>> property)
  {
    if (property.NodeType != ExpressionType.Lambda)
      throwEx();

    var body = property.Body;
    if (body.NodeType == ExpressionType.Convert)
      if (body is UnaryExpression unary)
        body = unary.Operand;
      else
        throwEx();

    if (!(body is MemberExpression member))
      throwEx();

    if (!(member.Member is PropertyInfo pi))
      throwEx();

    return pi;

    void throwEx() =>
      throw new NotSupportedException($"The expression '{property}' isn't supported.");
  }
}

Answer 5

What about a throw away IEqualityComparer generic class?

public class ThrowAwayEqualityComparer<T> : IEqualityComparer<T>
{
  Func<T, T, bool> comparer;

  public ThrowAwayEqualityComparer(Func<T, T, bool> comparer)   
  {
    this.comparer = comparer;
  }

  public bool Equals(T a, T b)
  {
    return comparer(a, b);
  }

  public int GetHashCode(T a)
  {
    return a.GetHashCode();
  }
}

So now you can use Distinct with a custom comparer.

var distinctImages = allImages.Distinct(
   new ThrowAwayEqualityComparer<GalleryImage>((a, b) => a.Key == b.Key));

You might be able to get away with the <GalleryImage>, but I'm not sure if the compiler could infer the type (don't have access to it right now.)

And in an additional extension method:

public static class IEnumerableExtensions
{
  public static IEnumerable<TValue> Distinct<TValue>(this IEnumerable<TValue> @this, Func<TValue, TValue, bool> comparer)
  {
    return @this.Distinct(new ThrowAwayEqualityComparer<TValue>(comparer);
  }

  private class ThrowAwayEqualityComparer...
}

Answer 6

This is currently in .NET 8 preview 2. The OP would use

var distinctImages = allImages.Distinct<GalleryImage>(
    EqualityComparer<GalleryImage>.Create((a, b) => a.id == b.id));

in place of their suggested

var distinctImages = allImages.Distinct<GalleryImage>(
    new EqualityComparer<GalleryImage>((a, b) => a.id == b.id));

More info:

https://github.com/dotnet/runtime/pull/75212

Answer 7

You could group by the key value and then select the top item from each group. Would that work for you?

Answer 8

Here's an interesting article that extends LINQ for this purpose... http://www.singingeels.com/Articles/Extending_LINQ__Specifying_a_Property_in_the_Distinct_Function.aspx

The default Distinct compares objects based on their hashcode - to easily make your objects work with Distinct, you could override the GetHashcode method.. but you mentioned that you are retrieving your objects from a web service, so you may not be able to do that in this case.

Answer 9

implement IEquatable on GalleryImage because it is generated

A different approach would be to generate GalleryImage as a partial class, and then have another file with the inheritance and IEquatable, Equals, GetHash implementation.