Python dictionary initialization and values append to list

Question: Write function that for a given list of objects, returns a dictionary of type: [object]. For example: func([1,2.0,3,'g']) returns: {'i':[1,3], 'f': [2.0], 'c': ['g'], 'o': []}

key i for ints
key f for floats
key c for chars
key o for others

My code:

def make_dict(lst):
    my_dict = {'i': [], 'f': [], 'c': [], 'o': []}
    [my_dict['i'].append(item) if isinstance(item, int) else my_dict['f'].append(item) if isinstance(
        item, float) else my_dict['c'].append(item) if isinstance(item, str) and len(item) == 1 else my_dict['o'].append(item) for item in lst]
    return my_dict

This function returning the corrent dictionary, but I need to make it a One-Liner, plus, I don't like the way it's done. I know the use of list comprehension is bad here.

My though process was to first make this a multiple lines function, which i did using for loops and if elif statements, Then i tried to combine those lines and eventually got it down to 3.

This is the for loop i used before combining it into a list comprehenson:

for item in lst:
    if isinstance(item, int):
        my_dict['i'].append(item)
    elif isinstance(item, float):
        my_dict['f'].append(item)
    elif isinstance(item, str) and len(item) == 1:
        my_dict['c'].append(item)
    else:
        my_dict['o'].append(item)
return my_dict

My problem: I know i need to use some sort of Dictionary comprehension combined with map() of filter() in some way, but after searching the internet and finding noting.. I just don't know how to do it.

Maybe you can help me? THANKS!

Solution

What you want is to group the data by the first letter of their type so

get the first letter of their type type(x).__name__[0]
group the data by this criteria

You need to provide data to groupby sorted by the same criteria you'll use for the grouping. the {k:list(v) part is to move from a groupby object to a dict

from itertools import groupby

def make_dict(lst):
    sorter = lambda x:type(x).__name__[0] if type(x).__name__ in ('int', 'float', 'str') else 'o'
    return {k:list(v) for k,v in groupby(sorted(lst, key=sorter), key=sorter)}

r = make_dict([1,2.0,3,'g']) # {'f': [2.0], 'i': [1, 3], 's': ['g']}
print(r)

Using filter, but not efficient as you iterate the list multiple times

def make_dict(lst):
    return {'i': list(filter(lambda x: isinstance(item, int), lst)), 
            'f': list(filter(lambda x: isinstance(item, float), lst)), 
            's': list(filter(lambda x: isinstance(item, str), lst)), 
            'o': type(x).__name__ not in ('int', 'float', 'str')}