Trying to understand how to create nested dictionaries on the fly. Ideally my dictionary would look something like:
mydict = { 'Message 114861156': { 'email': ['user1@domain.com', 'user2@domain.com'] }, { 'status': 'Queued mail for delivery' }}
Here's what i have so far:
sampledata = "Message 114861156 to user1@domain.com user2@domain.com [InternalId=260927844] Queued mail for delivery'."
makedict(sampledata)
def makedict(results):
newdict = {}
for item in results:
msgid = re.search(r'Message \d+', item)
msgid = msgid.group()
newdict[msgid]['emails'] = re.findall(r'\w+@\w+\.\w+', item)
newdict[msgid]['status'] = re.findall(r'Queued mail for delivery', item)
has the following output:
Traceback (most recent call last):
File "wildfires.py", line 57, in
striptheshit(q_result)
File "wildfires.py", line 47, in striptheshit
newdict[msgid]['emails'] = re.findall(r'\w+@\w+\.\w+', item)
KeyError: 'Message 114861156'
How do you make a nested dictionary like this on the fly?
解决方案
dict.setdefault is a good tool, so is collections.defaultdict
Your problem right now is that newdict is an empty dictionary, so newdict[msgid] refers to a non-existent key. This works when assigning things (newdict[msgid] = "foo"), however since newdict[msgid] isn't set to anything originally, when you try to index it you get a KeyError.
dict.setdefault lets you sidestep that by initially saying "If msgid exists in newdict, give me its value. If not, set its value to {} and give me that instead.
def makedict(results):
newdict = {}
for item in results:
msgid = re.search(r'Message \d+', item).group()
newdict.setdefault(msgid, {})['emails'] = ...
newdict[msgid]['status'] = ...
# Now you KNOW that newdict[msgid] is there, 'cuz you just created it if not!
Using collections.defaultdict saves you the step of calling dict.setdefault. A defaultdict is initialized with a function to call that produces a container that any non-existent key gets assigned as a value, e.g.
from collections import defaultdict
foo = defaultdict(list)
# foo is now a dictionary object whose every new key is `list()`
foo["bar"].append(1) # foo["bar"] becomes a list when it's called, so we can append immediately
You can use this to say "Hey if I talk to you about a new msgid, I want it to be a new dictionary.
from collections import defaultdict
def makedict(results):
newdict = defaultdict(dict)
for item in results:
msgid = re.search(r'Message \d+', item).group()
newdict[msgid]['emails'] = ...
newdict[msgid]['status'] = ...