IEquatable Interface what to do when checking for null

倖福魔咒の 提交于 2019-11-30 21:25:08

I've always found it easier to write the static operator with null handling, and have the Equals override call the overloaded operator with "this" as one of the parameters.

From Guidelines for Overloading Equals() and Operator == (C# Programming Guide)

//add this code to class ThreeDPoint as defined previously
//
public static bool operator ==(ThreeDPoint a, ThreeDPoint b)
{
    // If both are null, or both are same instance, return true.
    if (System.Object.ReferenceEquals(a, b))
    {
        return true;
    }

    // If one is null, but not both, return false.
    if (((object)a == null) || ((object)b == null))
    {
        return false;
    }

    // Return true if the fields match:
    return a.x == b.x && a.y == b.y && a.z == b.z;
}

public static bool operator !=(ThreeDPoint a, ThreeDPoint b)
{
    return !(a == b);
}

This is how ReSharper creates equality operators and implements IEquatable<T>, which I trust blindly, of course ;-)

public class ClauseBE : IEquatable<ClauseBE>
{
    private int _id;

    public bool Equals(ClauseBE other)
    {
        if (ReferenceEquals(null, other))
            return false;
        if (ReferenceEquals(this, other))
            return true;
        return other._id == this._id;
    }

    public override bool Equals(object obj)
    {
        if (ReferenceEquals(null, obj))
            return false;
        if (ReferenceEquals(this, obj))
            return true;
        if (obj.GetType() != typeof(ClauseBE))
            return false;
        return Equals((ClauseBE)obj);
    }

    public override int GetHashCode()
    {
        return this._id.GetHashCode();
    }

    public static bool operator ==(ClauseBE left, ClauseBE right)
    {
        return Equals(left, right);
    }

    public static bool operator !=(ClauseBE left, ClauseBE right)
    {
        return !Equals(left, right);
    }
}

Check for null and return false. Equals should always be false if one of the operands is null;

I think this is a bit less cumbersome than casting to Object before checking for null:

ReferenceEquals(a, null)

Other answers give good solutions to the general problem.

However, your own code can be simplified into a relatively simple solution ...

Firstly, at the start of your == operator you have this:

    // First test
    if (a as object == null && b as object == null)
    {
        return true;
    }

This qualifies as "working too hard".

If ClauseBE is a reference type, then you only need to compare with null - the "as object" is redundant; equally, if ClauseBE is a value type, then it can never be null.

Assuming that ClauseBE is a reference type (the most likely case), then you can simplify to this - note that we use Object.Equals() to avoid infinite recursion and a stack blowout.

    // First test
    if (Object.Equals(a, null) && Object.Equals(b, null))
    {
        return true;
    }

One useful shortcut is to use Object.ReferenceEquals() - which handles nulls for you.

So you could write this instead:

    // First test
    if (Object.ReferenceEquals(a, b))
    {
        return true;
    }

with the bonus that this also handles the case where a and b are the same exact object.

Once you get past the Object.ReferenceEquals() test, you know that a and b are different.

So your next test:

    // Second test
    if ((a as object == null && b as object != null)
        || (b as object == null && a as object != null))
    {
        return false;
    }

can be simplified - since you know that if a is null, b cannot be null, and so on.

    // Second test
    if (Object.Equals(a, null) || Object.Equals(b, null))
    {
        return false;
    }

If this test fails, then you know that a and b are different, and that neither is null. A good time to call your overridden Equals().

    // Use the implementation of Equals() for the rest
    return a.Equals(b as object);
public class Foo : IEquatable<Foo>
{
    public Int32 Id { get; set; }

    public override Int32 GetHashCode()
    {
        return this.Id.GetHashCode();
    }

    public override Boolean Equals(Object obj)
    {
        return !Object.ReferenceEquals(obj as Foo, null)
            && (this.Id == ((Foo)obj).Id);

        // Alternative casting to Object to use == operator.
        return ((Object)(obj as Foo) != null) && (this.Id == ((Foo)obj).Id);
    }

    public static Boolean operator ==(Foo a, Foo b)
    {
        return Object.Equals(a, b);
    }

    public static Boolean operator !=(Foo a, Foo b)
    {
        return !Object.Equals(a, b);
    }

    public Boolean Equals(Foo other)
    {
        return Object.Equals(this, other);
    }
}

I have used the following approach and it seemed to work well for me. Infact, Resharper suggests this approach.

public bool Equals(Foo pFoo)
{
        if (pFoo == null)
            return false;
        return (pFoo.Id == Id);
}

public override bool Equals(object obj)
{
        if (ReferenceEquals(obj, this))
            return true;

        return Equals(obj as Foo);
}
Emperor XLII

I prefer to perform all the comparison logic in the Equals(T) method, and leave the "if this or that is null, else ..." in operator overloads to the framework.

The only tricky thing about overriding operator overloads is that you can no longer use those operators in your Equals implementation, for example to compare with null. Instead, object.ReferenceEquals can be used to achieve the same effect.

Following the TwoDPoint example in the MSDN Guidelines for Overriding Equals() and Operator == article, this is the pattern I generate when implementing value equality for types:

public override bool Equals( object obj ) {
  // Note: For value types, would use:
  // return obj is TwoDPoint && this.Equals( (TwoDPoint)obj );
  return this.Equals( obj as TwoDPoint );
}

public bool Equals( TwoDPoint other ) {
  // Note: null check not needed for value types.
  return !object.ReferenceEquals( other, null )
      && EqualityComparer<int>.Default.Equals( this.X, other.X )
      && EqualityComparer<int>.Default.Equals( this.Y, other.Y );
}

public static bool operator ==( TwoDPoint left, TwoDPoint right ) {
  // System.Collections.Generic.EqualityComparer<T> will perform the null checks 
  //  on the operands, and will call the Equals overload if necessary.
  return EqualityComparer<TwoDPoint>.Default.Equals( left, right );
}

public static bool operator !=( TwoDPoint left, TwoDPoint right ) {
  return !EqualityComparer<TwoDPoint>.Default.Equals( left, right );
}

The form above is the safest implementation, as it simply forwards the field equality checks to the framework and requires no knowledge of whether the fields overload the equality operators. It is perfectly fine to simplify this where you know the overload exists:

public bool Equals( TwoDPoint other ) {
  return !object.ReferenceEquals( other, null )
      && this.X == other.X
      && this.Y == other.Y;
}

You can also replace the EqualityComparer<T> calls in the operator overloads with calls to the static object.Equals method when comparing reference types, or when boxing value types does not matter:

public static bool operator ==( TwoDPoint left, TwoDPoint right ) {
  return object.Equals( left, right );
}

public static bool operator !=( TwoDPoint left, TwoDPoint right ) {
  return !object.Equals( left, right );
}

See also What is the best algorithm for an overridden GetHashCode? for implementing GetHashCode.

标签
易学教程内所有资源均来自网络或用户发布的内容,如有违反法律规定的内容欢迎反馈
该文章没有解决你所遇到的问题?点击提问,说说你的问题,让更多的人一起探讨吧!