How to implement a good __hash__ function in python [duplicate]

Question

When implementing a class with multiple properties (like in the toy example below), what is the best way to handle hashing?

I guess that the __eq__ and __hash__ should be consistent, but how to implement a proper hash function that is capable of handling all the properties?

class AClass:
  def __init__(self):
      self.a = None
      self.b = None

  def __eq__(self, other):
      return other and self.a == other.a and self.b == other.b

  def __ne__(self, other):
    return not self.__eq__(other)

  def __hash__(self):
      return hash((self.a, self.b))

I read on this question that tuples are hashable, so I was wondering if something like the example above was sensible. Is it?

Answer 1

It's dangerous to write

def __eq__(self, other):
  return other and self.a == other.a and self.b == other.b

because if your rhs (i.e., other) object evaluates to boolean False, it will never compare as equal to anything!

In addition, you might want to double check if other belongs to the class or subclass of AClass. If it doesn't, you'll either get exception AttributeError or a false positive (if the other class happens to have the same-named attributes with matching values). So I would recommend to rewrite __eq__ as:

def __eq__(self, other):
  return isinstance(other, self.__class__) and self.a == other.a and self.b == other.b

If by any chance you want an unusually flexible comparison, which compares across unrelated classes as long as attributes match by name, you'd still want to at least avoid AttributeError and check that other doesn't have any additional attributes. How you do it depends on the situation (since there's no standard way to find all attributes of an object).

Answer 2

Documentation for object.__hash__(self)

The only required property is that objects which compare equal have the same hash value; it is advised to mix together the hash values of the components of the object that also play a part in comparison of objects by packing them into a tuple and hashing the tuple. Example

def __hash__(self):
    return hash((self.name, self.nick, self.color))