1

I have two inter-related questions about attributes and dir():

  1. What are attributes? Is it an object inside of another object? Or is it a method, property, etc?

  2. When you do dir(), what's the differences between the slots and the rest of the attributes?

0

1 Answer 1

2

Attributes

What are attributes? Is it an object inside of another object? Or is it a method, property, etc?

Semantically, in English, properties and attributes are the same thing. However, in Python, we typically mean property to mean an attribute that has its access (lookup, assignment, and deletion) programmatically managed.

Attributes provide information about objects. For example, a circle may be defined by its radius. Thus we might say the defining attribute of a circle is its radius attribute. We could model such an object like this:

class Circle(object):
    def __init__(self, radius):
        self.radius = radius

However, other information we might like to know about such a circle could be its diameter, circumference, or its area. These are also attributes of a circle. However, since we can calculate them from the radius, we might choose to use Python's property decorator:

    @property
    def diameter(self):
        return self.radius * 2

A property defined in this way would be read-only. We could do similarly for circumference and area (but the implementation is left as an exercise for the reader). We use both the radius attribute and the diameter property in the same way:

>>> c = Circle(5)
>>> c.radius
5
>>> c.diameter
10

But we will find that if we try to change the diameter, we cannot directly, we would have to change the radius instead.

>>> c.diameter = 20
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: can't set attribute
>>> c.radius = 10
>>> c.diameter
20

Everything is an object in Python

Is it an object inside of another object?

Well, in Python everything is an object, so all data members of an object are objects too. In objects that don't declare a __slots__, the data members are stored in a dictionary (another object) on the __dict__ attribute name.

If objects declare __slots__ the data members are stored in a kind of mutable tuple, but there are lots of caveats for __slots__.

Methods

Or is it a method, property, etc?

A method is actually more of an attribute of the class definition (it is in the class's __dict__, not the instance's), but it is bound to the instance on a dotted lookup from an instance (for more on this, see descriptors - an advanced topic). Semantically, it's also an attribute of the instance though.

You can read more about methods here.

Attributes and dir

When you do dir(), what's the differences between the slots and the rest of the attributes?

dir just tells you (most of) the names that are available on the object.

>>> dir(c)
['__class__', '__delattr__', ... 'diameter', 'radius'] # some removed

So that includes functions and properties which are actually attributes of the class of the object.

If you want to know what are the actual data members in an object, use the vars builtin:

>>> vars(c)
{'radius': 5}

And by this inspection we see that the data stored for this circle is not both radius and diameter, but only radius.

(A unique counter-example for dir: If you do dir(AClass) it won't tell you about the AClass.mro() though, which is mostly a CPython implementation detail.)

1
  • 1
    Amazing!! Thank you very much. So basically everything in class object is called attribute. You can define for example animal by it's skin(class property) and you can tell how this animal will act(class method), two of them is attributes. Am I right ? When you do dir() you can see also what you see in dict why is that? And cool! I know mro works with class method after that it goes to bases , and go on.. Thanks a lot!!
    – Adi Gal
    Mar 10, 2016 at 7:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.