CODEWITHSUNDEEP
pythonset
Ajayc
Ajayc

Reputation: 843

Set intersection of object dropping keys

I'm trying to perform some set operations using this object. 

def Group(object):
  def __init__(self, name, email, description, skip_if_deleted, membership=None)
    self.name = name
    self.email = remove_special_characters(email.lower(), bad_chars)
    self.description = description
    self.action = None
    self.skip_if_deleted = skip_if_deleted
    if membership is not None:
      self.membership = set(membership)
    else:
      self.membership = set([])

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

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

I am finding that when doing set intersections, I seem to be dropping the populated key 'membership', in exchange for an empty one.

An example:

Fetch from the first source, which does not provide membership information at time of fetch

# Example obj in unpopulated_membership looks like: Group(name, email, description, membership = empty set)
unpopulated_membership = set(get_groups_from_first_source())

Fetch from the second source, which does return membership information

# Example obj in populated_membershop looks like: Group(name, email, description, membership = populated set)
populated_membership = set(get_groups_from_other_source())

common_groups = populated_membership.intersection(unpopulated_membership)

Now at this point, common groups gives us the commonalities between the two sets, but the key 'membership' is now equal to unpopulated_membership's value. Regardless of whether i do populated_membership.intersection() or unpopulated_membership.intersection()

Why is this? How can i retain the membership information found in populated_membership when performing this intersection?

Thanks!

Upvotes: 0

Views: 32

Answers (1)

user2357112
user2357112

Reputation: 281330

The quick-fix would be to use a set comprehension:

common_groups = {group for group in populated_membership
                       if group in unpopulated_membership}

I would recommend choosing a different way to represent your data, though. For example, using dicts to match email addresses with the information associated with them. You've defined your Group.__hash__ and Group.__eq__ in such a way that "equal" Groups hold different data, which is highly unusual.

Also, if this is Python 2, don't forget to define __ne__. You should also define __eq__ and __ne__ in such a way that they return NotImplemented instead of throwing an AttributeError when the other object isn't a Group.

Upvotes: 1

Related Questions