How to flatten nested python dictionaries?

Question

I’m trying to flatten the nested dictionary:

dict1 = {
    'Bob': {
        'shepherd': [4, 6, 3],
        'collie': [23, 3, 45],
        'poodle': [2, 0, 6],
    },
    'Sarah': {
        'shepherd': [1, 2, 3],
        'collie': [3, 31, 4],
        'poodle': [21, 5, 6],
    },
    'Ann': {
        'shepherd': [4, 6, 3],
        'collie': [23, 3, 45],
        'poodle': [2, 10, 8],
    }
}

I’d like to display all values in lists:
[4, 6, 3, 23, 3, 45, 2, 0, 6, 1, 2, 3,…, 2, 10, 8]

My first idea was to do in this way:

dict_flatted = [ i for name in names.values() for dog in dogs.values() for i in dog]

Though I’m getting the error. I’d be happy for tips how to handle it.

Asked By: Monica

||

Source

Answer 1

You can use a simple recursive function as follows.

def flatten(d):    
    res = []  # Result list
    if isinstance(d, dict):
        for key, val in d.items():
            res.extend(flatten(val))
    elif isinstance(d, list):
        res = d        
    else:
        raise TypeError("Undefined type for flatten: %s"%type(d))

    return res


dict1 = {
    'Bob': {
        'shepherd': [4, 6, 3],
        'collie': [23, 3, 45],
        'poodle': [2, 0, 6],
    },
    'Sarah': {
        'shepherd': [1, 2, 3],
        'collie': [3, 31, 4],
        'poodle': [21, 5, 6],
    },
    'Ann': {
        'shepherd': [4, 6, 3],
        'collie': [23, 3, 45],
        'poodle': [2, 10, 8],
    }
}

print( flatten(dict1) )

Answered By: Daewon Lee

Answer 2

You’re trying to use variables that don’t exist.

Use this

dict_flatted = [ i for names in dict1.values() for dog in names.values() for i in dog]

Answered By: Philip Rollins

Answer 3

The recursive function can be single line:

def flatten(d):
    return d if isinstance(d, list) else sum((flatten(i) for i in d.values()), [])

Nota: was an answer for a duplicate question

Answered By: Serge Ballesta

Answer 4

I started with @DaewonLee’s code and extended it to more data types and to recurse within lists:

def flatten(d):
    res = []  # type:list  # Result list
    if isinstance(d, dict):
        for key, val in d.items():
            res.extend(flatten(val))
    elif isinstance(d, list):
        for val in d:
            res.extend(flatten(val))
    elif isinstance(d, (float, str, int)):
        res = [d]  # type: List[float] or List[str] or List[int]
    else:
        raise TypeError("Undefined type for flatten: %s" % type(d))
    return res

Answered By: Hephaestus

Answer 5

for this specific case: (2 levels)

using itertools.chain method:

nested_dict = {1 : {‘a’ : [1,2,3], ‘b’ : [4,5]},
2 : {‘a’ : [6], ‘b’ : [7,8,9]}}

list(chain(*[v for y in nested_dict.values() for v in list(y.values())]))

[1, 2, 3, 4, 5, 6, 7, 8, 9]

using pandas:

df = pd.DataFrame(pd.json_normalize(data=nested_dict), dtype=object)
df['chain'] = df.apply(lambda x: list(chain(*x.values.tolist())), axis=1)

Answered By: omer

How to flatten nested python dictionaries?

Question:

Answers: