Aggregate items in dict

Question

I have a list like this:

A = [{u'CI': {u'RP': 1}}, {u'CI': {u'RP': 1}},  {u'JI': {u'RP': 1}}]

and I want merge same keys and increment value in dict.

Example:

From these values :

{u'CI': {u'RP': 1}}, {u'CI': {u'RP': 1}}

I will have:

{u'CI': {u'RP': 2}}

final list result is:

A = [{u'CI': {u'RP': 2}}, {u'JI': {u'RP': 1}]

Answer 1

You can use a defaultdict from collections to help here. This is a dict that will create default values for missing keys. Firstly you'll want a defaultdict that has a default value of 0 to do your aggregation. Next you'll need a defaultdict that has the first kind of defaultdict as its default so you can build up to the two levels.

>>> A = [{u'CI': {u'RP': 1}}, {u'CI': {u'RP': 1}},  {u'JI': {u'RP': 1}}]
>>> B = defaultdict(lambda: defaultdict(int))
>>> for d in A:
...     for (key,d2) in d.iteritems():
...         for (key2, value) in d2.iteritems():
...             B[key][key2] += value
... 
>>> B.items()
[(u'CI', defaultdict(<type 'int'>, {u'RP': 2})), (u'JI', defaultdict(<type 'int'>, {u'RP': 1}))]

If you really want a list of single key dicts rather than a single dict with different keys you can convert the defaultdict as follows:

>>> [{key: dict(value)} for key,value in B.iteritems()]
[{u'CI': {u'RP': 2}}, {u'JI': {u'RP': 1}}]

Answer 2

Use the setdefault method of dict:

A = [{u'CI': {u'RP': 1}}, {u'CI': {u'RP': 1}},  {u'JI': {u'RP': 1}}]
B = {}
for i in A:
     k = i.keys()[0]             # k is 'CI' or 'JI' in this case
     B.setdefault(k, {u'RP': 0}) # set the default 'RP' to 0
     B[k]['RP'] += i[k][u'RP']   # add the RP. Because we already 
                                 # set the default to 0 this will not blow up
print B
# {u'CI': {u'RP': 2}, u'JI': {u'RP': 1}}