How do you implement a case-insensitive (or caseless) dictionary? In this article we explore solutions where we inherit from the built-in dict, the abstract base class MutableMapping from the standard module collections.abc, and the UserDict in the standard module collections.
In a recent article, I showed why you should use the string method casefold when working with case-insensitive strings.
In that article, I also presented a basic implementation of a case-insensitive dictionary.
However, that implementation (reproduced below) had a couple of issues.
class CaseInsensitiveDict(dict):
"""Case-insensitive dictionary implementation."""
def __getitem__(self, key):
return dict.__getitem__(self, key.casefold())
def __setitem__(self, key, value):
return dict.__setitem__(self, key.casefold(), value)
def __delitem__(self, key):
return dict.__delitem__(self, key.casefold())Our class CaseInsensitiveDict implements the three dunder methods __getitem__, __setitem__, and __delitem__, and that lets us mimic the main behaviour of dict.
However, this implementation is incomplete.
We can create an instance of a case-insensitive dictionary and expose the issues with the implementation above:
>>> d = CaseInsensitiveDict()
>>> d["mathspp"] = "https://mathspp.com/"
>>> d.get("MATHSPP", "not found")
'not found'The issues that we will find all revolve around what the example above shows. Although we implemented the three main dunder methods that let you add / access / delete keys from a dictionary, the other dictionary methods will not use the dunder methods to perform their operations.
dictIf we wanted to implement a case-insensitive dictionary that inherits from dict directly, we would have to provide implementations of most of the dict methods in terms of the three dunder methods that we already have.
If you are up for the challenge, open your Python REPL and inspect the result of dir(dict).
That list will contain all the (dunder) methods that dict knows about.
Go over that list and implement all the methods that need to be written in terms of our case-insensitive operations.
Such an implementation is included below for reference:
_no_default = object()
class CaseInsensitiveDict(dict):
"""Case-insensitive dictionary implementation."""
def __getitem__(self, key):
return dict.__getitem__(self, key.casefold())
def __setitem__(self, key, value):
return dict.__setitem__(self, key.casefold(), value)
def __delitem__(self, key):
return dict.__delitem__(self, key.casefold())
# ---
def __contains__(self, key):
return dict.__contains__(self, key.casefold())
def __init__(self, seed=None, **kwargs):
super().__init__()
# Defer work to the method .update.
self.update(seed)
self.update(kwargs)
def __or__(self, other):
"""Dictionary updating with the pipe operator |."""
base = self.copy()
base.update(other)
return base
def __ror__(self, other):
"""Dictionary updating with | and a regular dict on the left."""
base = CaseInsensitiveDict(other)
base.update(other)
return base
def copy(self):
return CaseInsensitiveDict(dict.copy(self))
def get(self, key, default=None):
return dict.get(self, key.casefold(), default)
def pop(self, key, default=_no_default):
if default is _no_default:
return dict.pop(self, key.casefold())
else:
return dict.pop(self, key.casefold(), default)
def setdefault(self, key, default=None):
return dict.setdefault(self, key.casefold(), default)
def update(self, seed=None, **kwargs):
if seed is None:
seed = {}
# Is the seed a mapping...
if hasattr(seed, "items"):
for key, value in seed.items():
self[key] = value
# or an iterable?
else:
for key, value in seed:
self[key] = value
for key, value in kwargs.items():
self[key] = valueStay tuned for the remainder of the article!
+35 chapters. +400 pages. Hundreds of examples. Over 30,000 readers!
My book “Pydon'ts” teaches you how to write elegant, expressive, and Pythonic code, to help you become a better developer. >>> Download it here 🐍🚀.
collections.abc.MutableMapping https://docs.python.org/3/library/collections.abc.html#collections.abc.MutableMapping [last accessed 29-01-2023];collections.UserDict https://docs.python.org/3/library/collections.html#userdict-objects [last accessed 29-01-2023];