Copying dictionaries in Python is easy, but with a few gotchas in special cases.
Dictionaries with immutable values can be copied simply by passing as arguments to the dict() constructor method. Immutable types in Python are int, float, decimal, bool, str, tuple, and range.
Example:
original_dict = dict({'foo': 'bar', 'fizz': ('fuzz', 'buzz')})
copied_dict = dict(original_dict)
That was easy. Now comes the more interesting part. In the case that the source dictionary has one or more mutable values, copying requires more caution depending on the desired behavior. Mutable types in Python are list, dictionary, set, and user-defined classes. Using the constructor dict() as above, will only produce a shallow copy of the source dictionary. This means that the corresponding values of each dictionary are the same object and have the same id. Changing the value of copied dict will be reflected in the original dict as well, since it’s the same underlying object being shared between the dictionaries.
original_dict = dict({'foo': ['bar', 'fuzz']})
shallow_copy = dict(original_dict)
id(original_dict['foo']) == id(shallow_copy['foo'])
In order to make a deep-copy of dictionaries with mutable values, Python provides a copy module which has the deepcopy method.
from copy import deepcopy
original_dict = dict({'foo': ['bar', 'fuzz']})
deep_copy = deepcopy(original_dict)
id(original_dict['foo']) == id(deep_copy['foo'])
Now you know how copying of dictionaries works in Python. 🙂