Python 3：按字段元组对 dict 列表进行分组

Question

I need to group a list of dict by a tuple of fields and write a function, which will do this for different sets of data.

For example: a list of dict:

[{'bnf_code': '0101010G0AAABAB',
 'items': 2,
  'practice': 'N81013',
  'bnf_name': 'Co-Magaldrox_Susp 195mg/220mg/5ml S/F',
  'nic': 5.98,
  'act_cost': 5.56,
  'quantity': 1000},
 {'bnf_code': '0101021B0AAAHAH',
  'items': 1,
  'practice': 'N81013',
  'bnf_name': 'Alginate_Raft-Forming Oral Susp S/F',
  'nic': 1.95,
  'act_cost': 1.82,
  'quantity': 500}]

and I want to groupby a tuple -> ('bnf_name', 'post_code')) .

I can't use pandas.

I've tried with below code but it raises a KeyError.

def group_by_field(data,fields):
    groups = {name: [] for name in fields}
    for script in data:
        for name in fields:
            groups[script[name]].append(script)
    return groups

Answer 1

Here is a function that does what you describe:

from collections import defaultdict
def group_by_field(data, fields):
    res = defaultdict(list)
    for d in data:
        key = tuple(d[f] for f in fields)
        res[key].append(d)
    return dict(res)