KeyError while defining a new key value pair in a defaultdict created from an existing dictionary

Question

Below is a sample code:

from collections import defaultdict

normal_dict = {
    "test": {
        "test1": 1,
        "test2": 2,
        "test3": {
            "test3_1": 1
        }
    }
}

default_dict = defaultdict(lambda: defaultdict(dict), normal_dict)

default_dict['test']['test4']['test4_1'] = 1 # Doesn't work


print(default_dict)

I get this error

Traceback (most recent call last):
  File "<string>", line 17, in <module>
KeyError: 'test4'

It seems like the previously non-existent key 'test4' is not being automatically added when I use the line default_dict['test']['test4']['test4_1'], for some reason. Theoretically, defaultdict should create the key test4 with another defaultdict as its default value.

The problem is that the defaultdict creation rules don't convert your fixed dictionaries from normal_dict into defaultdicts. If the initialization did try to coerce your input into its format then the ["test"]["test1"] would barf because 1 isn't a dict. I bet if you try you will see that default_dict['testX']['test4']['test4_1'] = 1 does work. And then if you just print(default_dict) you will see the difference between the 'test' entry and the 'testX' entry. — Malcolm, Jul 18 '22 at 21:26
@Malcolm That makes sense. I guess the question now is how would you convert all the dictionaries inside `normal_dict` into defaultdicts? — bravesheeptribe, Jul 18 '22 at 21:33
Does this answer your question? [Nested defaultdict of defaultdict](https://stackoverflow.com/questions/19189274/nested-defaultdict-of-defaultdict) — 0x5453, Jul 18 '22 at 21:46

Dani Mesejo · Accepted Answer · 2022-07-18T21:41:11.690

The Problem

The problem is that:

default_dict['test']

is:

{'test1': 1, 'test2': 2, 'test3': {'test3_1': 1}}

in other words is a dictionary not a defaultdict. Note that defaultdict provides a value (through the factory method) for a missing key, which is not the case at creation time in:

default_dict = defaultdict(lambda: defaultdict(dict), normal_dict)

It won't copy the values and "transform" them to defaultdict.

One solution

One approach, that transform every nested dictionary into a defaultdict is the following:

def nested_dd():
    # https://stackoverflow.com/a/19189356/4001592
    return defaultdict(nested_dd)


def convert_to_default(d):
    res = nested_dd()
    for key, value in d.items():
        if isinstance(value, dict):
            res[key] = convert_to_default(value)
        else:
            res[key] = value
    return res


result = convert_to_default(normal_dict)

result['test']['test4']['test4_1'] = 1
pprint.pprint(result)  # need to import pprint

Output

defaultdict(<function nested_dd at 0x7fd27008bd90>,
            {'test': defaultdict(<function nested_dd at 0x7fd27008bd90>,
                                 {'test1': 1,
                                  'test2': 2,
                                  'test3': defaultdict(<function nested_dd at 0x7fd27008bd90>,
                                                       {'test3_1': 1}),
                                  'test4': defaultdict(<function nested_dd at 0x7fd27008bd90>,
                                                       {'test4_1': 1})})})

Is it possible to convert all the nested dictionaries inside `normal_dict` into default dictionaries? — bravesheeptribe, Jul 18 '22 at 21:26

KeyError while defining a new key value pair in a defaultdict created from an existing dictionary

1 Answers1

The Problem

One solution