如何串联或合并两个具有重叠键的defaultdict的defaultdict？

Question

Working with the suggestion provided here , and having two revenue tables to have to extract from (containing the same fruits but different brands):

I end of with two defaultdicts of defaultdicts :

d1

defaultdict(<function <lambda> at 0x105dd42a8>, {u'Kiwi': defaultdict(<type 'int'>, {u'NZKiwi': 1.2}), u'Pear': defaultdict(<type 'int'>, {u'PearShaped': 6.2}), u'Banana': defaultdict(<type 'int'>, {u'BananaBrand': 4.0, u'OtherBrand': 3.2}), u'Apple': defaultdict(<type 'int'>, {u'CrunchApple': 1.7})})

d2

defaultdict(<function <lambda> at 0x105dd41b8>, {u'Kiwi': defaultdict(<type 'int'>, {u'n': 1.2}), u'Pear': defaultdict(<type 'int'>, {u'p': 6.2}), u'Banana': defaultdict(<type 'int'>, {u'b': 4.0, u'o': 3.2}), u'Apple': defaultdict(<type 'int'>, {u'a': 1.7})})

Assuming for argument's sake that I cannot combine the two tables before the above processing, how do I now combine these defaultdicts such that the keys (the fruit types) are maintained, yet their respective values get concatenated, per key? In other words, how do I end up with the following:

defaultdict(<function <lambda> at 0x105dd41b8>, {u'Kiwi': defaultdict(<type 'int'>, {u'n': 1.2, u'NZKiwi': 1.2}), u'Pear': defaultdict(<type 'int'>, {u'p': 6.2, u'PearShaped': 6.2}), u'Banana': defaultdict(<type 'int'>, {u'b': 4.0, u'o': 3.2, u'BananaBrand': 4.0, u'OtherBrand': 3.2}}), u'Apple': defaultdict(<type 'int'>, {u'a': 1.7, u'CrunchApple': 1.7})})

Actually it doesn't necessarily have to take that structure - really what I need is something from which I can extract the individual brand-amount pairs, per product . So end is results is a dict for Banana containing its 4 brand-amount pairs, another dict for Kiwi containing its two brand-amount pairs etc.

UPDATE: I could do the following, but there is surely a cleaner way??

BananaDict = {}
KiwiDict ={}
AppleDict = {}
PearDict = {}

for k, v in d1.iteritems():
    if k == 'Banana':
        BananaDict.update(v)
    elif k == 'Kiwi'
    etc.
    etc.

then repeat the above for d2.

Answer 1

You can just use a simple for loop , example -

>>> d = defaultdict(int, {u'Kiwi': defaultdict(int, {u'NZKiwi': 1.2}), u'Pear': defaultdict(int, {u'PearShaped': 6.2}), u'Banana': defaultdict(int, {u'BananaBrand': 4.0, u'OtherBrand': 3.2}), u'Apple': defaultdict(int, {u'CrunchApple': 1.7})})
>>> d1 = defaultdict(int, {u'Kiwi': defaultdict(int, {u'n': 1.2}), u'Pear': defaultdict(int, {u'p': 6.2}), u'Banana': defaultdict(int, {u'b': 4.0, u'o': 3.2}), u'Apple': defaultdict(int, {u'a': 1.7})})
>>> for k,v in d.items():
...     v.update(d1[k])
...
>>> d
defaultdict(<class 'int'>, {'Banana': defaultdict(<class 'int'>, {'o': 3.2, 'OtherBrand': 3.2, 'BananaBrand': 4.0, 'b': 4.0}), 'Pear': defaultdict(<class 'int'>, {'PearShaped': 6.2, 'p': 6.2}), 'Kiwi': defaultdict(<class 'int'>, {'NZKiwi': 1.2, 'n': 1.2}), 'Apple': defaultdict(<class 'int'>, {'a': 1.7, 'CrunchApple': 1.7})})

Answer 2

I noticed that the other answer didnt account for the case where the new dictionary had a key that the original dictionary didnt have.

def merge_defaultdicts(d,d1):
    for k,v in d1.items():
        if (k in d):
            d[k].update(d1[k])
        else:
            d[k] = d1[k]
    return d