扁平化词典列表

Question

I have a list of dictionaries, which may or may not have similar keys, and I want to flatten the list into a single dictionary, with a list type for values.

An example of this as follows:

data =  [{'category': u'Non-profit organization', 'categories': [u'Theater',
      u'Bar', u'Concert Venue']}, {'category': u'Non-profit organization', 
      'categories': [u'Business Services', u'College & University']}]

This should become the following:

print result
result = {'category': [u'Non-profit organization', u'Non-profit 
      organization'], 'categories': [u'Theater', u'Bar', u'Concert Venue',       
      u'Business Services', u'College & University']]}

As you can see, anything that is a string value in the initial data should be added as a value within a list. Anything that is held in list type in the initial dictionary should be added it its key, but to create a flattened list.

Clearly a simple solution is to for loop through it all and append values, but am looking for a simpler solution.

Answer 1

Use a defaultdict(list) :

from collections import defaultdict

res = defaultdict(list)
for dic in data:
    for key, value in dic.items():
        old_value = res[key]
        if isinstance(value, list):
            old_value.extend(value)
        else:
            old_value.append(value)

        # alternatively:
        old_value += [value] if not isinstance(value, list) else value

The reasoning is: in the end you want all values to be list s. The difference is that values that originally where lists should be joined together (and list.extend does that) while other values should be inserted in a new list (as in list.append ).

---

Also, there is no built-in method nor class in the collections module that does this automatically, so I believe the above is probably "optimal" as far as code dimension, readability and efficiency are considered.

Answer 2

In additional previous answer (sorry, I haven't enough reputation to write a comment) you need to check if value in res otherwise you will have repeat values:

{'category': [u'Non-profit organization', u'Non-profit organization'], 'categories': [u'Theater', u'Bar', u'Concert Venue', u'Business Services', u'College & University']}

twice u'Non-profit organization'