counting common key-value pairs in a list of dictionaries

Question

Suppose I have a list of dictionaries that all share the keys color and object :

inpt = [{'object':'square', 'color':'red', 'size':'big'},
        {'object':'square', 'color':'red', 'coord':(0,0)},
        {'object':'square', 'color':'red'},
        {'object':'triangle', 'color':'blue', 'adj':'beautiful'},
        {'object':'triangle', 'color':'blue', 'attr':'none'}]

Here I only care about object and color . I'd like to count the number of red squares and blue triangles for instance. This means finding the number of times {'object':'square', 'color':'red'} and {'object':'triangle', 'color':'blue'} occurred. In other words, I need to find the number of commonalities in the key-value pairs in a list of dictionaries like this.

The end result could for example be something like this:

{('square', 'red'): 3, ('triangle', 'blue'):2}

What is an easy way to do this?

Answer 1

Just use a collections.Counter() object , feeding it tuples of the values:

from collections import Counter

result = Counter((d['object'], d['color']) for d in inpt)

This gives you a dictionary subclass with (object, color) keys. You could get a list of tuples, in descending count order, using the Counter.most_common() method :

result = result.most_common()

Demo:

>>> from collections import Counter
>>> inpt = [{'object':'square', 'color':'red', 'size':'big'},
...         {'object':'square', 'color':'red', 'coord':(0,0)},
...         {'object':'square', 'color':'red'},
...         {'object':'triangle', 'color':'blue', 'adj':'beautiful'},
...         {'object':'triangle', 'color':'blue', 'attr':'none'}]
>>> Counter((d['object'], d['color']) for d in inpt)
Counter({('square', 'red'): 3, ('triangle', 'blue'): 2})
>>> _.most_common()
[(('square', 'red'), 3), (('triangle', 'blue'), 2)]