使用Python中的词典列表创建新词典

Question

I have a list of dictionaries, where some values are strings, and other values are integers:

list_countries = [{'country' : 'Suriname', 
            'population' : 532724,
            'capital': 'Paramaribo', 
            'anthem': 'God zij met ons Suriname'},
            {'country' : 'Sweden', 
            'population' : 9683248,
            'capital': 'Stockholm', 
            'anthem': 'Du gamla, Du fria'},
             ...]

I would like to reshape each of these key-value pairs into one big new dictionary. However, my approach has issues:

dict_countries = { 'countries':       [],
                   'pop':             [],
                   'capital_city':    [],
                   'national_anthem': [] }

I then iterate through and append all values with .extend() .

for dictionary in list_countries:
    dict_countries['countries'].extend(dictionary['country'])
    dict_countries['pop'].extend(dictionary['population'])
    dict_countries['capital_city'].extend(dictionary['capital'])
    dict_countries['national_anthem'].extend(dictionary['anthem'])

However, this doesn't work. All strings are broken up, letter by letter. For the integers, I get the error:

TypeError: 'int' object is not iterable

What is the correct approach to do this?

EDIT: I believe there is a value for each key. However, let's say there isn't. How would I rewrite the above in order to add NaN if there is no value found.

Answer 1

.extend() expects the argument passed to it to be an iterable, ie. a list or a string etc. As per your example, population is an integer and therefore not iterable hence your exception message.

If you change it to .append() it will behave as you expect it to.

Answer 2

The reason you are getting the output you are is because there is a difference between append and extend on a list. If you extend with an iterable as a parameter (which a string is) it would inline each item of the iterable into the dict (which is each letter for the string). However it fails for the int as that is not an iterable. I would prefer to use append which simply appends to the list in the dict.

list_countries = [{'country' : 'Suriname', 
            'population' : 532724,
            'capital': 'Paramaribo', 
            'anthem': 'God zij met ons Suriname'},
            {'country' : 'Sweden', 
            'population' : 9683248,
            'capital': 'Stockholm', 
            'anthem': 'Du gamla, Du fria'}]

dict_countries = { 'countries':       [],
                   'pop':             [],
                   'capital_city':    [],
                   'national_anthem': [] }

for dictionary in list_countries:
    dict_countries['countries'].append(dictionary['country'])
    dict_countries['pop'].append(dictionary['population'])
    dict_countries['capital_city'].append(dictionary['capital'])
    dict_countries['national_anthem'].append(dictionary['anthem'])

print dict_countries

Answer 3

You have two problems to solve here:
Build lists of the values from your dictionaries and translate the old keys to their new names.

Use the setdefault method of the builtin dictionary and use a translation-dictionary as a dictionary in the literal sense (ie for translation).

Set up the translations like this:

>>> translations = {'country': 'countries',
...                 'population': 'pop',
...                 'capital': 'capital_city',
...                 'anthem': 'national_anthem'}

Then build your new dictionary:

>>> merged = {}
>>> for d in list_countries:
...     for k in d:
...         key = translations.get(k, k)
...         merged.setdefault(key, []).append(d[k])
... 
>>> merged
{'national_anthem': ['God zij met ons Suriname', 'Du gamla, Du fria'], 'capital_city': ['Paramaribo', 'Stockholm'], 'pop': [532724, 9683248], 'countries': ['Suriname', 'Sweden']}

... and if you can be sure all dictionaries share the same keys, here's a oneliner:

>>> {translations.get(k,k):[d[k] for d in list_countries] for k in list_countries[0].keys()}
{'national_anthem': ['God zij met ons Suriname', 'Du gamla, Du fria'], 'capital_city': ['Paramaribo', 'Stockholm'], 'pop': [532724, 9683248], 'countries': ['Suriname', 'Sweden']}

Answer 4

Here's what I've done. The only limitation is that the keys in the new dictionary are not grammatically pluralised but I guess you can do that manually at the end.

list_countries = [{'country' : 'Suriname', 
            'population' : 532724,
            'capital': 'Paramaribo', 
            'anthem': 'God zij met ons Suriname'},
            {'country' : 'Sweden', 
            'population' : 9683248,
            'capital': 'Stockholm', 
            'anthem': 'Du gamla, Du fria'},
             ]


from collections import defaultdict
d = defaultdict(list)
for i in list_countries:
   for k,v in i.items():
     d[k].append(v)

The d can be coverted back into a regular dict easily.

Answer 5

keys =  list_countries[0].keys()
values = (list(t) for t in zip(*[d.values() for d in list_countries]))
dict(zip(keys, values))