I was playing around with Dictionary and stumbled across the below scenario
public class MyObject
{
public string I { get; set; }
public string J { get; set; }
public string K { get; set; }
public override int GetHashCode()
{
int hashCode = (I+J+K).GetHashCode();
Debugger.Log(9, "INFO", hashCode.ToString() + System.Environment.NewLine);
return hashCode;
}
}
class Program
{
static void Main(string[] args)
{
MyObject obj1 = new MyObject() { I = "Hello", J = "World" };
MyObject obj2 = new MyObject() { I = "Hello", J = "World" };
Dictionary<MyObject, string> collection = new Dictionary<MyObject, string>();
collection.Add(obj1, "1");
var result = collection[obj2]; // KeyNotFound exception here.
}
}
I have MyObject class which acts as a key to dictionary and I override the GetHashCode method to return hash code based on the values stored in the class.
So, when the above code is executed, both obj1 and obj2 returns same hash code, but still the dictionary throws KeyNotFound exception.
Any reason why such a behavior?
In .NET,
GetHashCodeis used in concert with theEqualsmethod to determine object equality with regard to storage in collections.Note that a hash table is more complex than simply mapping a key to a single slot via a hash code. Due to the nature of hash codes, collisions can occur and do occur in practice (though with a good hash function this should not be very often). Thus most hash table implementations have to deal with the case of two different objects generating the same hash code and this is often achieved with a linked list at each “slot” in the hash table. The hash code is used to determine the slot and the
Equalsmethod is used to determine whereabouts the object is stored in the linked list (in most “standard” implementations of a hash table).A word of warning, however: there are very few good reasons to override the built-in behaviour of
GetHashCode. I found this interesting SO thread discussingGetHashCodeandEqualsthat should be worth a read: Why is it important to override GetHashCode when Equals method is overridden?. It discusses the merit/demerits of changing the behaviour, the properties of good and bad hash functions, the required properties of these two methods and other goodies.