add object into python's set collection and determine by object's attribute

Question

I have a Person class like this:

class Person:
    def __init__(self, name, age):
        self.name = name
        self.age = age

    def __repr__(self):
        return '<Person {}>'.format(self.name)

I want to add some instances of this class to a set, like this:

tom = Person('tom', 18)
mary = Person('mary', 22)
mary2 = Person('mary2', 22)

person_set = {tom, mary, mary2}
print(person_set)
# output: {<Person tom>, <Person mary>, <Person mary2>}

As you can see, there are 2 Marys in the set. How can I make it so that Person instances with the same age are considered the same person, and only added to the set once?

In other words, how can I get a result of {<Person tom>, <Person mary>}?

So, as an example, `Frank, age 19` would be the same as `Heather, age 19` if you are just comparing by age? — jamylak, May 11 '12 at 07:41
@sashimi, then you need a dictionary with age as key, Person object as value. — spinlok, May 11 '12 at 07:48

score 41 · Accepted Answer · edited Jul 26 '18 at 21:37

41

When a new object is being added to a python set, the hash code of the object is first computed and then, if one or more objects with the same hash code is/are already in the set, these objects are tested for equality with the new object.

The upshot of this is that you need to implement the __hash__(...) and __eq__(...) methods on your class. For example:

class Person:
    def __init__(self, name, age):
        self.name = name
        self.age = age

    def __eq__(self, other):
        return self.age == other.age

    def __hash__(self):
        return hash(self.age)

    def __repr__(self):
        return '<Person {}>'.format(self.name)

tom = Person('tom', 18)
mary = Person('mary', 22)
mary2 = Person('mary2', 22)

person_set = {tom, mary, mary2}
print(person_set)
# output: {<Person tom>, <Person mary>}

However, you should think very carefully about what the correct implementation of __hash__ and __eq__ should be for your class. The above example works, but is non-sensical (e.g. in that both __hash__ and __eq__ are defined only in terms of age).

edited Jul 26 '18 at 21:37

Aran-Fey

39,665
11
104
149

answered May 11 '12 at 07:57

srgerg

18,719
4
57
39

thanks~this is what I want,I know it is weird.the class like **Person** which I will use is very simple.so it doesn't matter wheather overwrite or not.I know it's undesirable ^_^,I will use it carefully – sashimi May 11 '12 at 08:18
1

Why does python even require specifying the `__eq__` method, if it's already doing the comparison of the value specified in `__hash__` ? – wesinat0r Jun 16 '20 at 15:48
1

I answered my own comment - https://docs.python.org/3/glossary.html#term-hashable `__hash__` is for the lifetime value of the "object", `__eq__` is for comparing with other objects. You still need both to do comparison – wesinat0r Jun 16 '20 at 18:02
1

https://hynek.me/articles/hashes-and-equality/ is a good article for more background – wesinat0r Jun 17 '20 at 18:48

add object into python's set collection and determine by object's attribute

1 Answers1

Linked

Related