Overriding GetHashCode

Question

As you know, GetHashCode returns a semi-unique value that can be used to identify an object instance in a collection. As a good practice, it is recommended to override this method and implement your own.

My question is - do you override this method when working on custom objects? If so, what algorithm do you use to generate the unique ID?

I was thinking about generating a GUID and then getting integer data from that identificator.

Have a read of [this question](http://stackoverflow.com/questions/263400/what-is-the-best-algorithm-for-an-overridden-system-object-gethashcode) and its answers. It outlines a good hash code implementation. There is also a good discussion on overiding `GetHashCode` for mutable objects [here](http://stackoverflow.com/questions/873654/overriding-gethashcode-for-mutable-objects-c). — adrianbanks, Nov 21 '10 at 21:35
I don't know what "semi-unique" means... a value is either unique or it's not, and a hashcode is not unique. Therefore it doesn't allow you to identify an object in a list. And it's not "a good practice" to override `GetHashCode`, that's something you do when you *need* to (e.g. to use the object as a key in a dictionary), not because you think it's a good practice. — Thomas Levesque, Nov 21 '10 at 21:38
Using the object as a key can be considered as identifying the object in a collection - that's exactly why I'm looking for input on what is the best algorithm to build the identifier. As of semi-unique IDs: http://www.west-wind.com/Weblog/posts/4741.aspx — Den, Nov 21 '10 at 22:03
Dennis, you first need to think what equality behaviour you want. For the default reference-equality (for mutable objects with identity) you have to do exactly nothing. — H H, Nov 21 '10 at 22:11

score 30 · Answer 1 · edited Apr 28 '16 at 17:22

30

If you use resharper it can generate the GetHashCode(), Equals and operator method bodies for you.

Access this menu by pressing Alt+Insert.

http://www.jetbrains.com/resharper/webhelp/Code_Generation__Equality_Members.html

edited Apr 28 '16 at 17:22

RubberDuck

11,933
4
50
95

answered Dec 02 '10 at 16:52

Winston

500
3
3

score 28 · Accepted Answer · edited Nov 24 '20 at 13:45

28

When you override GetHashCode() you also need to override Equals(), operator== and operator!= . And be very careful to meet all the requirements for those methods.

The guidelines are here on MSDN. Most important quote:

It is not a good idea to override operator == in mutable types.

edited Nov 24 '20 at 13:45

Michael Puckett II

6,586
5
26
46

answered Nov 21 '10 at 22:07

H H

263,252
30
330
514

From what I understand, that will make no sense because, well, the object won't change it's state. – Den Nov 21 '10 at 22:18
What will make no sense? – H H Nov 21 '10 at 22:19
3

strings are immutable, have you never used `name == "x"` ? – H H Nov 21 '10 at 22:22
Yes, but when you modify a string, you create a new object (e.g. when you use Replace). – Den Nov 21 '10 at 22:26
Still, string overrides Equals and all the rest. I'm not getting your point here. – H H Nov 21 '10 at 22:29
@Dennis: You cannot modify a string, it is *immutable*. That is not a semantic statement. Calling methods like Replace which return a new string does not modify a string, not literally, and not figuratively. You have to change your mentality here. – Tergiver Nov 21 '10 at 23:45
Oh no, I am not saying that string is mutable. I was saying exactly what you just stated - when I call Replace, the string is not modified - a new object is created instead. – Den Nov 22 '10 at 00:30
3

The point is that it's natural to rely on == staying true until you reassign the variable; it is not intuitive to have mutation side effects change the past meaning of == as anything that supports comparison operators by default are value types or immutable. These only change value when you say var = "new value" which explicitly implies a previous == may not be true anymore. References types can change state from other code and methods mutate the object elsewhere could change the equality. Equals does not provide the same expectation because it is a method definition, not syntax. – TheXenocide Jun 21 '12 at 16:32
Just to clarify: Non-immutable is the same thing as mutable? – jasdefer Jan 20 '17 at 09:22
@jasdefer - yes, but mutable is kind of the norm in imperative languages like C#. – H H Jan 20 '17 at 09:38

score 1 · Answer 3 · answered Nov 21 '10 at 21:33

In my personal usage, I only override when overriding equals method too. Generally, I do this for objects I know that I might run a LINQ to Objects query on, or some other comparison operation.

I usually return, if say a LINQ to SQL entity or DTO object, the primary key value. Whatever you return, if you don't store the value locally, it may produce an unexpected result.

HTH.

score 1 · Answer 4 · edited May 23 '17 at 12:10

I would normally override hashcode and equality checking methods for data classes (i.e. classes where the value semantics makes sense). Have a look at this question for a common implementation. If you do override hashcode override equals. Using a GUID is a pretty terrible idea because you want two objects which are different instances but have the same value to have the same hashcode and for equals to return true.

score 0 · Answer 5 · edited May 23 '17 at 10:31

0

you only need to override GetHashCode if you are overriding Equals. The default GetHashCode is implemented by the runtime in a similar way you wanted to do it - every object has a hidden field assigned by the runtime.

How to override GetHashCode

Actually your IDE should do this for you - when you type "override GetHashCode" the IDE should generate this boilerplate code. Visual Studio does not do it but SharpDevelop does.

edited May 23 '17 at 10:31

Community

1
1

answered Nov 21 '10 at 21:41

Martin Konicek

39,126
20
90
98

The default implementation of the GetHashCode method does not guarantee unique return values for different objects. Consequently, the default implementation of this method must not be used as a unique object identifier for hashing purposes. From: http://msdn.microsoft.com/en-us/library/system.object.gethashcode.aspx – brain Nov 21 '10 at 21:45
in VS one has to use the `equals` snippet instead – CodesInChaos Nov 21 '10 at 22:16
1

@brain - I read this. Does it mean that you should always override GetHashCode when you want to store your objects in a Dictionary? I don't think so - almost noone does this. The default implementation works fine for purposes of Dictionary. It is not perfect, but fine: http://stackoverflow.com/questions/750947/net-unique-object-identifier – Martin Konicek Dec 11 '10 at 14:58

score -2 · Answer 6 · answered Nov 21 '10 at 21:30

-2

Generally I use the aggregated GetHashCode from the component properties of the class. E.g.

public class Test
{
  public string Text { get; set; }
  public int Age { get; set; }

  public override GetHashCode()
  {
    int result = 
      string.IsNullOrEmpty(Text) ? 0 : Text.GetHashCode()
      + Age.GetHashCode();

    return result;
  }
}

answered Nov 21 '10 at 21:30

Matthew Abbott

60,571
9
104
129

5

This is not a very goo way to do it, see http://stackoverflow.com/questions/263400/what-is-the-best-algorithm-for-an-overridden-system-object-gethashcode – Martin Konicek Nov 21 '10 at 21:43
In general adding isn't good for creating a compound hashcode. But in this specific example I see no problem with it. But of course the overriding GetHashCode without overriding Equals doesn't make much sense. – CodesInChaos Nov 21 '10 at 22:19
Well, you learn something every day! – Matthew Abbott Nov 21 '10 at 22:28

Overriding GetHashCode

6 Answers6

Linked