Reputation: 545
I am modifying somebody's code in the context of the 'gym' environment and came across the use of dot notation to access a dictionary. The following snippet shows that the dictionary in 'gym' can use the notation, but when I duplicate it, it throws an error.
import gym
env = gym.Env
env = make('connectx', debug=True)
config = env.configuration
print(config)
print(config.timeout)
dct = {'timeout': 5, 'columns': 7, 'rows': 6, 'inarow': 4, 'steps': 1000}
print(dct.timeout)
this provides the following output:
{'timeout': 5, 'columns': 7, 'rows': 6, 'inarow': 4, 'steps': 1000}
5
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
<ipython-input-45-674d59d34c55> in <module>
6 print(config.timeout)
7 dct = {'timeout': 5, 'columns': 7, 'rows': 6, 'inarow': 4, 'steps': 1000}
----> 8 print(dct.timeout)
AttributeError: 'dict' object has no attribute 'timeout'
I am using Python 3. Can somebody explain please? Thanks
Upvotes: 4
Views: 10493
Reputation: 2285
You can create a custom dictionary class that supports dot notation. You can subclass the built-in dict
class and override the __getattr__
, __setattr__
, and __delattr__
methods. This will allow you to set, access, and delete dictionary items while retaining the usual dictionary behavior:
class DotDict(dict):
def __getattr__(self, key):
try:
return self[key]
except KeyError:
raise AttributeError(f"'DotDict' object has no attribute '{key}'")
def __setattr__(self, key, value):
self[key] = value
def __delattr__(self, key):
try:
del self[key]
except KeyError:
raise AttributeError(f"'DotDict' object has no attribute '{key}'")
my_dict = DotDict(name='John', age=30)
print(my_dict) # {'name': 'John', 'age': 30}
print(my_dict.name) # John
# Delete a key with dot notation
del my_dict.age
print(my_dict) # {'name': 'John'}
# Create a dictionary
new_dict = DotDict({'name': 'John', 'age': 30})
print(new_dict) # {'name': 'John', 'age': 30}
We can extend this for nested dictionaries even further by modifying the DotDict
class to recursively convert nested dictionaries into DotDict
instances:
class DotDict(dict):
def __init__(self, *args, **kwargs):
super(DotDict, self).__init__(*args, **kwargs)
for key, value in self.items():
if isinstance(value, dict):
self[key] = DotDict(value)
def __getattr__(self, key):
try:
return self[key]
except KeyError:
raise AttributeError(f"'DotDict' object has no attribute '{key}'")
def __setattr__(self, key, value):
self[key] = value
def __delattr__(self, key):
try:
del self[key]
except KeyError:
raise AttributeError(f"'DotDict' object has no attribute '{key}'")
my_dict = DotDict({'user': {'name': 'John', 'birth': {'month': 'june', 'day': 8}}})
print(my_dict.user.birth) # {'month': 'june', 'day': 8}
print(my_dict.user.birth.month) # june
I published the package dotmapdict which simplifies the whole process. You can install it with:
pip install dotmapdict
Upvotes: 0
Reputation: 1444
Unlike JavaScript objects, Python dicts don't natively support dot notation.
Try Dotsi, Addict or a similar library. Here's a quick snippet of Dotsi in action:
>>> import dotsi
>>>
>>> d = dotsi.Dict({"foo": {"bar": "baz"}}) # Basic
>>> d.foo.bar
'baz'
>>> d.users = [{"id": 0, "name": "Alice"}] # In list
>>> d.users[0].name
'Alice'
>>>
Disclosure: I'm Dotsi's author.
Upvotes: 5
Reputation: 2113
In python you cannot access a dictionnary value with dict.key
, you need to use dict[key]
Example :
d = {"foo": 2}
print(d["foo"])
# 2
key = foo
print(d[key])
# 2
print(d.foo)
# AttributeError: 'dict' object has no attribute 'foo'
print(d.key)
# AttributeError: 'dict' object has no attribute 'key'
If you really want to use the dot notation, you can use a class (your config
is probably a class instance by the way):
class MyClass():
def __init__(self):
self.foo = "bar"
a = MyClass()
print(a.foo)
# bar
Upvotes: 2