如何根据组合键值总结字典值？

Question

My question is similar to this( Python sum on keys for List of Dictionaries ), but need to sum up the values based on two or more key-value elements. I have a list of dictionaries as following:

list_to_sum=
        [{'Name': 'A', 'City': 'W','amt':100},
         {'Name': 'B', 'City': 'A','amt':200},
         {'Name': 'A', 'City': 'W','amt':300},
         {'Name': 'C', 'City': 'X','amt':400},
         {'Name': 'C', 'City': 'X','amt':500},
         {'Name': 'A', 'City': 'W','amt':600}]
            

So based on a combination of Name and City key values, amt should be summed. Please let me know how to solve this.

Output: [{'Name': 'A', 'City': 'W','amt':900},
         {'Name': 'B', 'City': 'A','amt':200},
         {'Name': 'C', 'City': 'X','amt':900}]

Answer 1

You could create a collections.Counter .Then you can simply add the values as the appear using the tuple as (Name, City) as the key:

from collections import Counter

list_to_sum=[
    {'Name': 'A', 'City': 'W','amt':100},
    {'Name': 'B', 'City': 'A','amt':200},
    {'Name': 'A', 'City': 'W','amt':300},
    {'Name': 'C', 'City': 'X','amt':400},
    {'Name': 'C', 'City': 'X','amt':500},
    {'Name': 'A', 'City': 'W','amt':600}
]
    
totals = Counter()

for d in list_to_sum:
    totals[(d['Name'],d['City'])] += d['amt']

print(totals[('A','W')]) # 1000
print(totals[('B','A')]) # 200
print(totals[('C','X')]) # 900
      

This will produce a dictionary-like object Counter :

Counter({('A', 'W'): 1000, ('B', 'A'): 200, ('C', 'X'): 900})

With this you can convert the dict back to a list of dicts like:

sums_list = [{'Name':Name, 'City':City, 'amt':amt} for (Name, City), amt in totals.items()]

giving sums_list :

[{'Name': 'A', 'City': 'W', 'amt': 1000},
 {'Name': 'B', 'City': 'A', 'amt': 200},
 {'Name': 'C', 'City': 'X', 'amt': 900}]

Answer 2

list_to_sum = [{'Name': 'A', 'City': 'W', 'amt': 100},
               {'Name': 'B', 'City': 'A', 'amt': 200},
               {'Name': 'A', 'City': 'W', 'amt': 300},
               {'Name': 'C', 'City': 'X', 'amt': 400},
               {'Name': 'C', 'City': 'X', 'amt': 500},
               {'Name': 'A', 'City': 'W', 'amt': 600}]

sum_store = {}

for entry in list_to_sum:
    key = (entry['Name'], entry['City'])
    if key in sum_store:
        sum_store[key] += entry['amt']
    else:
        sum_store[key] = entry['amt']

print(sum_store)

output:

{('A', 'W'): 1000, ('B', 'A'): 200, ('C', 'X'): 900}

Answer 3

Besides the answers proposed by the others, it can be done in a pandas one-liner. It groups rows by name and city and calculates the sum over their amt feature.

import pandas as pd
list_to_sum=[
    {'Name': 'A', 'City': 'W','amt':100},
    {'Name': 'B', 'City': 'A','amt':200},
    {'Name': 'A', 'City': 'W','amt':300},
    {'Name': 'C', 'City': 'X','amt':400},
    {'Name': 'C', 'City': 'X','amt':500},
    {'Name': 'A', 'City': 'W','amt':600}
]

df = pd.DataFrame(list_to_sum)

t = df.groupby(['Name','City']).amt.sum()

print(t)

Output:
Name  City
A     W       400
B     A       200
C     X       900