Creating a unique list of dictionaries from a list of dictionaries which contains same keys but different values

Question

suppose i have been given a list of dictionaries like this :

[{"id":1, "symbol":'A', "num":4}, {"id":2, "symbol":'A', "num":3}, {"id":1, "symbol":'A', "num":5}, {"id":2, "symbol":'B', "num":1}]

Now, i have to create a dictionary or alter the current one such that (id, symbol) together are unique and the num value is the sum of all the values present in the dict with that (id,symbol) so that the new dict or the current dict looks something like this:

[{"id":1, "symbol":'A', "num":9}, {"id":2, "symbol":'A', "num":3}, {"id":2, "symbol":'B', "num":1}]

Answer 1

I would go the pandas.DataFrame way -

import pandas as pd

a = [{"id":1, "symbol":'A', "num":4}, {"id":2, "symbol":'A', "num":3}, {"id":1, "symbol":'A', "num":5}, {"id":2, "symbol":'B', "num":1}]
b = pd.DataFrame(a)
c = b.groupby(['id', 'symbol'])['num'].sum().reset_index()
print(c)
d = list(c.to_dict(orient='index').values())
print(d)

Output

[{'id': 1, 'symbol': 'A', 'num': 9}, {'id': 2, 'symbol': 'A', 'num': 3}, {'id': 2, 'symbol': 'B', 'num': 1}]

Hope that works!

Answer 2

You can use groupby in the following manner:

from itertools import groupby
from operator import itemgetter

grouper = itemgetter("id", "symbol")
result = []
for key, grp in groupby(sorted(input_data, key = grouper), grouper):
    temp_dict = dict(zip(["id", "symbol"], key))
    temp_dict["num"] = sum(item["num"] for item in grp)
    result.append(temp_dict)

from pprint import pprint
pprint(result)

Output:

[{'id': 1, 'num': 9, 'symbol': 'A'},
 {'id': 2, 'num': 3, 'symbol': 'A'},
 {'id': 2, 'num': 1, 'symbol': 'B'}]

Answer 3

Here's a solution with plain python.

dlst = [{"id":1, "symbol":'A', "num":4}, {"id":2, "symbol":'A', "num":3},
    {"id":1, "symbol":'A', "num":5}, {"id":2, "symbol":'B', "num":1}]


# Create a dict where keys are tuples of (id,symbol), values are num

combined_d = {}

for d in dlst:
    id_sym = (d["id"], d["symbol"])
    if id_sym in combined_d:
        combined_d[id_sym] += d["num"]
    else:
        combined_d[id_sym] = d["num"]

# create a list of dictionaries from the tuple-keyed dict

result = []

for k, v in combined_d.items():
    d = {"id": k[0], "symbol": k[1], "num": v}
    result.append(d)

print(result)

It does what you want, but the resulting list is not sorted, as it's built from a dictionary.