[PYTHON] dict in dict Makes a dict a dict

Regarding the subject, I will report that it was implemented by a python beginner. In the first place, a dict in dict is a dict like the one below.

{'name': '2M2VgARX', 'js': {'a': 0, 'b': 0, 'c': 'sdmo'}}

The in dict key and the in dict key are unique even if they are combined, so I wanted to combine them into one dictionary, so I created such a function.

Image after conversion

{'name': '2M2VgARX', 'js_a': 0, 'js_b': 0, 'js_c': 'sdmo'}

There seems to be a standard function, but my search ability is low, so Please note that it was not found.

Implementation

import secrets
import string


def dict_extract(item, prefix=''):
    return_dict = {}
    for key, value in item.items():
        if isinstance(value, dict):
            prefix += f'{key}_'
            return_dict.update(dict_extract(value, prefix))
        else:
            return_dict[prefix + key] = value
    return return_dict


def main():
    # create sample data
    items =[]
    alphabet = string.ascii_letters + string.digits
    password = ''.join(secrets.choice(alphabet) for i in range(8))
    for i in range(10):
        password = ''.join(secrets.choice(alphabet) for i in range(8))
        tmp = {'a': i, 'b': i*2, 'c': 'sdmo'}
        tmp_item = {'name': password, 'js': tmp}
        items.append(tmp_item)

    # print dict
    for item in items:
        print(dict_extract(item))


if __name__ == "__main__":
    main()

Execution result

{'name': 'ZQ4NaBbc', 'js_a': 0, 'js_b': 0, 'js_c': 'sdmo'}
{'name': 'E5hS6IqS', 'js_a': 1, 'js_b': 2, 'js_c': 'sdmo'}
{'name': 'PC1xzQGi', 'js_a': 2, 'js_b': 4, 'js_c': 'sdmo'}
{'name': 'McJTGCjl', 'js_a': 3, 'js_b': 6, 'js_c': 'sdmo'}
{'name': '3hYsdwB8', 'js_a': 4, 'js_b': 8, 'js_c': 'sdmo'}
{'name': 'UKocBVhJ', 'js_a': 5, 'js_b': 10, 'js_c': 'sdmo'}
{'name': '3QVlhwOn', 'js_a': 6, 'js_b': 12, 'js_c': 'sdmo'}
{'name': 'bspVE4Hp', 'js_a': 7, 'js_b': 14, 'js_c': 'sdmo'}
{'name': 'rLLUxF4f', 'js_a': 8, 'js_b': 16, 'js_c': 'sdmo'}
{'name': '5wt3UCvR', 'js_a': 9, 'js_b': 18, 'js_c': 'sdmo'}

Commentary

• main()
  • # create sample data
    I'm just making a dict in dict, so I'll omit it. secrets seem to be safer than random. Official → https://docs.python.org/ja/3/library/secrets.html
  • # print dict
    Output for confirmation.
• dict_extract(item, prefix='') --The body of the implementation. --I made it a recursive function in consideration of dict in dict in dict ... --If the type of value is dict, try dict_extract again with value. --I tried connecting the keys with an underscore so that the key of the dict to be finally returned is unique. --If it is confirmed that it is unique, can I remove the `prefix + = f'{key} _' `part?

Digression

I feel that it is smarter to join with prefix as list. In that case, avoid passing list as an argument. list is a mutable, updatable object. If you specify a mutable object as the initial value of the argument, that object is created when the function is defined, so if you call the function with the corresponding argument omitted, the same object will be used.

code

def string_arg2(s_arg=['hatsumi']):
    s_arg.append(' san')
    print(''.join(s_arg))

for i in range(10):
    string_arg2()

result

hatsumi san
hatsumi san san
hatsumi san san san
hatsumi san san san san
hatsumi san san san san san
hatsumi san san san san san san
hatsumi san san san san san san san
hatsumi san san san san san san san san
hatsumi san san san san san san san san san
hatsumi san san san san san san san san san san

The result is like a business card management service. By the way, dict is also mutable, so be careful. It takes a lot of work to initialize.

def string_arg2(s_arg=None):
    if s_arg is None:
        s_arg = ['hatsumi']
    s_arg.append(' san')
    print(''.join(s_arg))

for i in range(10):
    string_arg2()

Recruitment

--There seems to be a standard function of the above implementation. I would be grateful if you could tell me how to find it. .. ――Is there a general name for dict in dict? Nested dictionary? It becomes d in d for short, and it looks like docker in docker and is confusing ...