group by property

[scidam]

Pandas is a very good tool to perform grouping, e.g.

Show the data

data = [
  {
    "name": "John",
    "m": "good",
    "n": 1
  },
  {
    "name": 'Alina',
    "m": "good",
    "n": 1
  },
  {
    "name": "Olivia",
    "m": "bad",
    "n": 2
  },
  {
    "name": "Ruby",
    "m": "bad",
    "n": 2
  },
    {
    "name": "Ruby",
    "m": "bad",
    "n": 3
  }
]

import pandas as pd
df = pd.DataFrame(data)
df.groupby(['m', 'n']).groups

Output:
{('bad', 2): Int64Index([2, 3], dtype='int64'),
 ('bad', 3): Int64Index([4], dtype='int64'),
 ('good', 1): Int64Index([0, 1], dtype='int64')}

If you don't want to use Pandas, you can do this using standard Python packages only, e.g.

from itertools import groupby
from operator import itemgetter
for a, b in groupby(data, key=lambda x: (itemgetter('m')(x), itemgetter('n')(x))):
    print(a, list(b))

Output:
('good', 1) [{'name': 'John', 'm': 'good', 'n': 1}, {'name': 'Alina', 'm': 'good', 'n': 1}]
('bad', 2) [{'name': 'Olivia', 'm': 'bad', 'n': 2}, {'name': 'Ruby', 'm': 'bad', 'n': 2}]
('bad', 3) [{'name': 'Ruby', 'm': 'bad', 'n': 3}]

This might be not exactly you want. Nevertheless, these approaches may be useful: you can easily access/process the data in each group.