Question

The MSDN documentation on Object.GetHashCode() describes 3 contradicting rules for how the method should work.

1. If two objects of the same type represent the same value, the hash function must return the same constant value for either object.
2. For the best performance, a hash function must generate a random distribution for all input.
3. The hash function must return exactly the same value regardless of any changes that are made to the object.

Rules 1 & 3 are contradictory to me.

Does Object.GetHashCode() return a unique number based on the value of an object, or the reference to the object. If I override the method I can choose what to use, but I’d like to know what is used internally if anyone knows.

Answer 1

Rules 1 & 3 are contradictory to me.

To a certain extent, they are. The reason is: if an object is stored in a hash table and, by changing its value, you change its hash then the hash table has lost the value and you can’t find it again by querying the hash table. — It is therefore important that while objects are stored in a hash table, they retain their hash value.

To ensure this it is often simplest to make hashable objects immutable, thus side-stepping the problem wholly. But it is actually sufficient to make only those fields immutable that determine the hash value.

Consider the following example:

struct Person {     public readonly string FirstName;     public readonly string Name;     public readonly DateTime Birthday;      public int ShoeSize; } 

People rarely change their name, and even less frequently their birthday. However, their shoe size may grow arbitrarily, or even shrink. It is therefore reasonable to identify people using their birthday and name but not their shoe size. The hash value should reflect this:

public override int GetHashCode() {     return HashCode.Combine(FirstName, Name, Birthday); } 

And remember to override Equals whenever you override GetHashCode (and vice-versa)!