Get all parents keys in nested dictionary for all items

Question

I want to get all parent keys for all items in a nested python dictionary with unlimited levels. Take an analogy, if you think of a nested dictionary as a directory containing sub-directories, the behaviour I want is similar to what glob.glob(dir, recursive=True) does.

For example, suppose we have the following dictionary:

sample_dict = {
  "key_1": {
    "sub_key_1": 1,
    "sub_key_2": 2,
  },
  "key_2": {
    "sub_key_1": 3,
    "sub_key_2": {
      "sub_sub_key_1": 4,
    },
  },
}

I want to get the full "path" of every value in the dictionary:

["key_1", "sub_key_1", 1]
["key_1", "sub_key_2", 2]
["key_2", "sub_key_1", 3]
["key_2", "sub_key_2", "sub_sub_key_1", 4]

Just wondering if there is a clean way to do that?

Answer 1

Using generators can often simplify the code for these type of tasks and make them much more readable while avoiding passing explicit state arguments to the function. You get a generator instead of a list, but this is a good thing because you can evaluate lazily if you want to. For example:

def getpaths(d):
    if not isinstance(d, dict):
        yield [d]
    else:
        yield from ([k] + w for k, v in d.items() for w in getpaths(v))

result = list(getpaths(sample_dict))

Result will be:

[['key_1', 'sub_key_1', 1],
 ['key_1', 'sub_key_2', 2],
 ['key_2', 'sub_key_1', 3],
 ['key_2', 'sub_key_2', 'sub_sub_key_1', 4]]

Answer 2

You can solve it recursively

sample_dict = {
  "key_1": {
    "sub_key_1": 1,
    "sub_key_2": 2,
  },
  "key_2": {
    "sub_key_1": 3,
    "sub_key_2": {
      "sub_sub_key_1": 4,
    },
  }
}

def full_paths(sample_dict, paths=[], parent_keys=[]):
    for key in sample_dict.keys():
        if type(sample_dict[key]) is dict:
            full_paths(sample_dict[key], paths=paths, parent_keys=(parent_keys + [key]))
        else:
            paths.append(parent_keys + [key] + [sample_dict[key]])
    return paths

print(full_paths(sample_dict))

Answer 3

You can use this solution.

sample_dict = {
  "key_1": {
    "sub_key_1": 1,
    "sub_key_2": 2,
  },
  "key_2": {
    "sub_key_1": 3,
    "sub_key_2": {
      "sub_sub_key_1": 4,
    },
  },
}

def key_find(sample_dict, li=[]):
    for key, val in sample_dict.items():
        if isinstance(val, dict):
            key_find(val, li=li + [key])
        else:
            print(li + [key] + [val])

key_find(sample_dict)