{"id":2575,"date":"2023-05-12T17:15:06","date_gmt":"2023-05-12T22:15:06","guid":{"rendered":"http:\/\/www.ishygddt.xyz\/~blog\/?p=2575"},"modified":"2023-05-26T15:20:18","modified_gmt":"2023-05-26T20:20:18","slug":"frozenset-as-frozendict","status":"publish","type":"post","link":"http:\/\/www.ishygddt.xyz\/~blog\/2023\/05\/frozenset-as-frozendict","title":{"rendered":"the real frozendict was the standard library we neglected along the way"},"content":{"rendered":"

If you need a quick immutable representation of dictionaries, but can't be bothered with a 3rd-party library, why not try frozenset(dict_items)<\/code>?<\/p>\n

It is immutable. It is in the standard library. It is unordered, so you're guaranteed<\/strong> that frozenset([('a', 1), ('b', 2)]) == frozenset([('b', 2), ('a', 1)])<\/code>. And it is the case that, for dictionaries with all immutable values (which is implied by the problem statement anyway), dict(frozenset(d.items())) == d<\/code>.<\/p>\n

Now, yes, it is possible to create<\/em> a frozenset<\/code> that isn't a valid representation of a dictionary (e.g. if not all(len(item) == 2 for item in something_which_isnt_dict_items)<\/code>). But that won't happen by calling frozenset(dict_items)<\/code>. \ud83d\ude42<\/p>\n

\n

And if the above really<\/em> isn't fancy enough for you, maybe try this on for size:<\/p>\n

import collections.abc\n\nclass frozenmapping(frozenset, collections.abc.Mapping):\n\tdef __iter__(self):\n\t\tyield from (k for k, v in super().__iter__())\n\tdef __getitem__(self, key):\n\t\tmissing = object()\n\t\tvalue = next((v for k, v in super().__iter__() if k == key), missing)\n\t\tif value is missing:\n\t\t\traise KeyError(key)\n\t\treturn value\n\tdef __contains__(self, key):\n\t\t# skip frozenset and force use of abc.Mapping's mixin\n\t\treturn super(frozenset, self).__contains__(key)\n\tdef __eq__(self, other):\n\t\t# skip frozenset and force use of abc.Mapping's mixin\n\t\treturn super(frozenset, self).__eq__(other)\n\tdef __neq__(self, other):\n\t\t# skip frozenset and force use of abc.Mapping's mixin\n\t\treturn super(frozenset, self).__neq__(other)\n\tdef __new__(cls, mapping_or_iterable):\n\t\tif isinstance(mapping_or_iterable, collections.abc.Mapping):\n\t\t\titems = mapping_or_iterable.items()\n\t\telse:\n\t\t\titems = mapping_or_iterable\n\t\treturn super().__new__(cls, ((k, v) for k, v in items))\n\tdef __repr__(self):\n\t\tif not self:\n\t\t\treturn f'{self.__class__.__name__}()'\n\t\td = dict(self)\n\t\treturn f'{self.__class__.__name__}({d!r})'\n\t# We didn't "have" to implement this; abc.Mapping would "have our back", kinda\n\t# but we will implement it to avoid O(N^2) performance\n\tdef items(self):\n\t\tyield from super().__iter__()<\/code><\/pre>\n
The handy thing about collections.abc.Mapping<\/code><\/a> is that all<\/strong> you need to provide is __getitem__<\/code>, __iter__<\/code>, and __len__<\/code>, and that class will magically \"mixin\" all the remaining mapping methods (__contains__<\/code>, keys<\/code>, items<\/code>, values<\/code>, get<\/code>, __eq__<\/code>, and __ne__<\/code>) that you didn't implement to create a fully featured dict<\/code>-a-like class. (Actually, it's not fully<\/em> featured since this one's not mutable\u2014obviously<\/em>, per our goal today\u2014if you want a mutable mapping instead, look into collections.abc.MutableMapping<\/code><\/a>.)<\/p>\n
Going over each of the methods:<\/p>\n