递归展平对象

Question

I have the following object:

d={
   'CollectionTitleIDs': {
           '852070#TVEpisode': '12', 
           '852063#TVEpisode': '4', 
           '852067#TVEpisode': '9'
     }, 
    'ReleaseYear': '2005', 
    'TVSeriesID': '5638#TVSeries'
}

I would like to flatten it to get the following output:

{
   'CollectionTitleIDs': [
      {"_Key": "852070#TVEpisode", "Value": "12"},
      {"_Key": "852063#TVEpisode", "Value": "4"},
      {"_Key": "852067#TVEpisode", "Value": "9"}
    ]
    'ReleaseYear': '2005', 
    'TVSeriesID': '5638#TVSeries'
}

In other words, if the value of the key is a dict, to push the keys of that dict to a "_Key" field and the value of that to a "Value" field.

I've been having difficulty with the recursion. And currently I have something like this:

What I currently have is:

def flatten_obj(obj, FLAT_OBJ=None):

    if FLAT_OBJ is None: FLAT_OBJ = OrderedDict()
    if isinstance(obj, list):
        return [flatten_obj(l, FLAT_OBJ=FLAT_OBJ) for l in obj]
    elif not isinstance(obj, dict):
        return obj
    else:
        for key in list(obj.keys()):
            val = get_sub_object_from_path(obj, key)
            if isinstance(val, dict):
                FLAT_OBJ[key] = [{'_Key': subkey, 'Value': flatten_obj(subval)} for subkey, subval in val.items()]
            elif isinstance(val, list):
                FLAT_OBJ[key] = flatten_obj(val, FLAT_OBJ=FLAT_OBJ)
            else:
                FLAT_OBJ[key] = val
    return FLAT_OBJ

Note that the above has one level of recursion, but it's possible the "value" will have a(nother) nested object as well, in which case we want to again extract the _Key,Value.

Answer 1

You can use unpacking with a list comprehension:

d = {'CollectionTitleIDs': {'852070#TVEpisode': '12', '852063#TVEpisode': '4', '852067#TVEpisode': '9'}, 'ReleaseYear': '2005', 'TVSeriesID': '5638#TVSeries'}
new_d = {**d, 'CollectionTitleIDs':[{'_Key':a, 'Value':b} for a, b in d['CollectionTitleIDs'].items()]}

Output:

{'CollectionTitleIDs': 
  [{'_Key': '852070#TVEpisode', 'Value': '12'}, 
   {'_Key': '852063#TVEpisode', 'Value': '4'}, 
   {'_Key': '852067#TVEpisode', 'Value': '9'}], 
 'ReleaseYear': '2005', 'TVSeriesID': '5638#TVSeries'}

Answer 2

The code should be simpler than that:

1. If it's a list return the list of flattened objects
2. If it's a dict return a list of the key/val objects
3. otherwise just return x

in code:

def flatten1(x):
    if isinstance(x, list):
        return [flatten1(y) for y in x]
    elif isinstance(x, dict):
        return [{"_Key": key, "Value": flatten1(value)}
                for key, value in x.items()]
    else:
        return x

In your example however for some reason the toplevel is not being flattened and is instead a regular dict, then just do so

def flatten(x):
    if isinstance(x, dict):
       return {key: flatten1(value) for key, value in x.items()}
    else:
       return flatten1(x)

flatten1 could even be made an internal function:

def flatten(x):
    def flatten1(x):
        if isinstance(x, list):
            return [flatten1(y) for y in x]
        elif isinstance(x, dict):
            return [{"_Key": key, "Value": flatten1(value)}
                    for key, value in x.items()]
        else:
            return x

    if isinstance(x, dict):
       return {key: flatten1(value) for key, value in x.items()}
    else:
       return flatten1(x)

Answer 3

My Method:

d={
   'CollectionTitleIDs': {
           '852070#TVEpisode': '12', 
           '852063#TVEpisode': '4', 
           '852067#TVEpisode': '9'
     }, 
    'ReleaseYear': '2005', 
    'TVSeriesID': '5638#TVSeries'
}

d['CollectionTitleIDs'] = list(map(lambda x: {'_Key': x[0], 'Value': x[1]}, d['CollectionTitleIDs'].items()))

#{
#  'CollectionTitleIDs': [
#    {'_Key': '852070#TVEpisode', 'Value': '12'},
#    {'_Key': '852063#TVEpisode', 'Value': '4'}, 
#    {'_Key': '852067#TVEpisode', 'Value': '9'}
#  ], 
#  'ReleaseYear': '2005', 
#  'TVSeriesID': '5638#TVSeries'
#}