[SOLVED] I'm getting a list of lists in my reducer output rather than a paired value and I am unsure of what to change in my code

I'm getting a list of lists in my reducer output rather than a paired value and I am unsure of what to change in my code

The code below is giving me nearly the output i want but not quite.

def reducer(self, year, words):
        x = Counter(words)
        most_common = x.most_common(3) 
        sorted(x, key=x.get, reverse=True)    
        yield (year, most_common)

This is giving me output

"2020" [["coronavirus",4],["economy",2],["china",2]]

What I would like it to give me is

"2020" "coronavirus china economy"

If someone could explain to me why i am getting a list of lists instead of the output i require I would be most grateful. Along with an idea on how to improve the code to get what I need.

Solution

From the documentation for Counter.most_common explains why you get a list of lists.

most_common(n=None) method of collections.Counter instance
    List the n most common elements and their counts from the most
    common to the least.  If n is None, then list all element counts.
    
    >>> Counter('abracadabra').most_common(3)
    [('a', 5), ('b', 2), ('r', 2)]

Because sorting from highest to lowest frequency is like sorting in descending order, but sorting alphabetically is in ascending order, you can use a custom tuple where you take the negative of the frequency and sort everything in ascending order.

from collections import Counter

words = Counter(['coronavirus'] * 4 + ['economy'] * 2 + ['china'] * 2 + ['whatever'])
x = Counter(words)
most_common = x.most_common(3)
# After sorting you need to discard the freqency from each (word, freq) tuple
result = ' '.join(word for word, _ in sorted(most_common, key=lambda x: (-x[1], x[0])))