How do I dynamically create properties in Python?

Question

Suppose I have a class like this:

class Alphabet(object):
     __init__(self):
         self.__dict = {'a': 1, 'b': 2, ... 'z': 26}

     @property
     def a(self):
         return self.__dict['a']

     @property
     def b(self):
         return self.__dict['b']

     ...

     @property
     def z(self)
         return self.__dict['z']

This would be a long and cumbersome task to define and it seems highly redundant. Is there a way to dynamically create these properties? I know you can dynamically create attributes with the built-in setattr, but I want to have control over read/write/delete access (for this reason I want to use a property). Thanks!

doing @property\ndef something() is equivalent to something = property(something). Which is equivalent to doing (in this case) self.setattr('a', property(lambda self: self.__dict['a'])) — Joel Cornett
– Joel Cornett, Commented Jun 10, 2012 at 9:29
@JoelCornett: The second statement in your comment is incorrect, you need to set it on the class, not the instance: paste.aeum.net/show/106. The reason for this is because of how descriptors work: The following methods only apply when an instance of the class containing the method (a so-called descriptor class) appears in an owner class (the descriptor must be in either the owner’s class dictionary or in the class dictionary for one of its parents). — ThiefMaster
– ThiefMaster, Commented Jun 10, 2012 at 9:35

ThiefMaster · Accepted Answer · 2012-06-10 09:39:57Z

45

Don't use properties but implement the following methods:

__getattr__(self, name)
__setattr__(self, name, value)
__delattr__(self, name)

See http://docs.python.org/reference/datamodel.html#customizing-attribute-access

Your __getattr__ method could look like this:

def __getattr__(self, name):
    try:
        return self.__dict[name]
    except KeyError:
        msg = "'{0}' object has no attribute '{1}'"
        raise AttributeError(msg.format(type(self).__name__, name))

edited Jun 10, 2012 at 9:39

answered Jun 10, 2012 at 9:26

ThiefMaster

320k85 gold badges608 silver badges648 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

jackberry Over a year ago

I guess, this loses the ability for IDEs to provide auto-complete attributes.

Katriel · Accepted Answer · 2012-06-10 10:07:42Z

-2

Don't do that. Just let the consumers of the class at the __dict directly and trust them not to screw it up. Remember, we're all consenting adults here!

Ned Batchelder explains better than I can:

Keep data out of your variable names

The question reminded me of others I've seen on Stack Overflow or in the #python IRC channel:

How do I see if a variable exists?

How do I use a variable as the name of another variable?

How do I use a variable as part of a SQL table name?

The thing all these have in common is trying to bridge the gap between two domains: the data in your program, and the names of data in your program. Any time this happens, it's a clear sign that you need to move up a level in your data modeling. Instead of 26 lists, you need one dictionary. Instead of N tables, you should have one table, with one more column in it.

answered Jun 10, 2012 at 10:07

Katriel

124k19 gold badges141 silver badges172 bronze badges

3 Comments

Karl Knechtel Over a year ago

Ned has a good point, but there is a goal here of encapsulation. That's not just important for preventing mistakes (which consenting adults can make, too), but as a matter of convenience. The internal code of the class probably has a good reason to treat the dictionary as a dictionary while client code treats it as a set of attributes.

martineau Over a year ago

A problem with accessing __dict directly is it's ugly, especially with the name mangling that will be involved. A little syntactic sugar for clients goes a long way sometimes, not to mention the level of indirection it provides will facilitate possible future changes.

error 1044 Over a year ago

in regards to @martineau's comment, the name mangling would make this solution verry ugly, this is because the dict is prefixed with a double underscore. Also it is an unspoken rule that in python anything prefixed with an underscore should not be touched. Also if you use a property to define a class attribute then you don't want it do be modified.

Collectives™ on Stack Overflow

How do I dynamically create properties in Python?

2 Answers 2

1 Comment

Keep data out of your variable names

3 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

1 Comment

3 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related