python - Programmatically accessing arbitrarily deeply-nested values in a dictionary -

this question has answer here:

• checking dictionary using dot notation string 4 answers

i'm working on python script i'm given lists of strings of format: ['key1', 'key2', 'key2.key21.key211', 'key2.key22', 'key3'].

each value in list corresponds entry in dictionary, , entries structured 'key2.key21.key211', correspond (in example), key 'key211' nested within 'key21', nested inside 'key2'.

the above list corresponds dictionary:

x = {      'key1' : 'value1',      'key2' : {                'key21' : {                           'key211': 'value211'                          },                'key22' : 'value22'               },      'key3' : 'value3'     } 

the names not regular key(n)+; can of form food.vegetables.potato, example. guarantees have key names themselves, in dictionary, not contain . character, , dictionary contains entries referenced in original list.

my question is, given such list of strings, how programmatically access corresponding entries in dictionary? can think of solution using eval(), 1 using traversal/search, want avoid calls eval(), , impression traversal comparisons slow (since dicts aren't search trees), , entail lot of nasty exception handling.

one approach write function access keys in nested dicts.

def deep_access(x,keylist):      val = x      key in keylist:          val = val[key]      return val  s = 'key2.key21.key211'  print deep_access(x,s.split('.')) 

result:

value211 

---

another approach, if want use similar syntax normal dictionary access subclass dict , override __getitem__ allow nested access when tuple of keys provided:

class nesteddict(dict):      def __getitem__(self,keytuple):         # if key not tuple access normal         if not isinstance(keytuple, tuple):             return super(nesteddict,self).__getitem__(keytuple)         d = self         key in keytuple:             d = d[key]         return d  >>> nd = nesteddict(x) >>> nd['key2'] {'key22': 'value22', 'key21': {'key211': 'value211'}} >>> nd['key2','key22'] 'value22' >>> nd['key2','key21'] {'key211': 'value211'} >>> nd['key2','key21','key211'] 'value211' 

you can implement __setitem__ , __delitem__ needed.