pythonic way to collect specific data from complex dict

i need to collect some data from complext dict based on Dot Notation key names

for example

sample data

data = {
  'name': {'last': 'smith', 'first': 'bob'},
  'address':{'city': 'NY', 'state': 'NY'}开发者_Python百科,
  'contact':{'phone':{'self':'1234', 'home':'222'}},
  'age':38,
  'other':'etc'
}

keys = ['contact.phone.self', 'name.last', 'age']

my logic

result = []
for rev_key in rev_keys:  
  current = data.copy()   
  rev_key = rev_key.split('.')  
  while rev_key:  
    value = rev_key.pop(0)  
    current = current[value]  
  result.append(current)

Thanks in advance!

[reduce(dict.get, key.split("."), data) for key in keys]

How about this?

def fetch( some_dict, key_iter ):
    for key in key_iter:
        subdict= some_dict
        for field in key.split('.'):
            subdict = subdict[field]
        yield subdict

a_dict = {
  'name': {'last': 'smith', 'first': 'bob'},
  'address':{'city': 'NY', 'state': 'NY'},
  'contact':{'phone':{'self':'1234', 'home':'222'}},
  'age':38,
  'other':'etc'
}

keys = ['contact.phone.self', 'name.last', 'age']

result = list( fetch( a_dict, keys ) )

Here's my crack at it:

>>> def find(tree,cur):
    if len(cur)==1:
        return tree[cur[0]]
    else:
        return find(tree[cur[0]],cur[1:])


>>> print [find(data,k.split(".")) for k in keys]
['1234', 'smith', 38]

Of course this will cause stack overflows if the items are too deeply nested (unless you explicitly raise the recursion depth), and I would use a deque instead of a list if this were production code.

Just write a function that gets one key at a time

def getdottedkey(data, dottedkey):
    for key in dottedkey.split('.'):
        data = data[key]
    return data

print [getdottedkey(data, k) for k in keys]