Search code examples
pythonsqlsortingpython-itertools

Sort and Flatten List/Dictionary of Data


I feel like this has to be fairly easy but find myself nesting too many loops for sanity.

Given a list of dictionaries like this

[{'id': 101, 'value1': 'yes', 'value2': '',    'value3': '',    'value4': 'no' },
 {'id': 102, 'value1': '',    'value2': 'yes', 'value3': '',    'value4': 'no'},
 {'id': 103, 'value1': 'no',  'value2': '',    'value3': 'yes', 'value4': ''   },
 {'id': 104, 'value1': 'yes', 'value2': '',    'value3': 'no',  'value4': 'yes'},
 {'id': 105, 'value1': '',    'value2': 'yes', 'value3': 'yes', 'value4': ''   },
 {'id': 106, 'value1': 'yes', 'value2': '',    'value3': '',    'value4': ''   }]

I want to create lists of dictionaries where the values either match or are empty.

Desired result (lists of lists but could be a list/dict/whatever, note there are duplicates)

[[{'id': 101, 'value1': 'yes', 'value2': '',    'value3': '',    'value4': 'no' },
  {'id': 102, 'value1': '',    'value2': 'yes', 'value3': '',    'value4': 'no' },
  {'id': 105, 'value1': '',    'value2': 'yes', 'value3': 'yes', 'value4': ''   },
  {'id': 106, 'value1': 'yes', 'value2': '',    'value3': '',    'value4': ''   }],

 [{'id': 102, 'value1': '',    'value2': 'yes', 'value3': '',    'value4': 'no' },
  {'id': 103, 'value1': 'no',  'value2': '',    'value3': 'yes', 'value4': ''   },
  {'id': 105, 'value1': '',    'value2': 'yes', 'value3': 'yes', 'value4': ''   }],

 [{'id': 104, 'value1': 'yes', 'value2': '',    'value3': 'yes', 'value4': 'yes'},
  {'id': 106, 'value1': 'yes', 'value2': '',    'value3': '',    'value4': ''   }]]

There has to be some way to do this using itertools groupby I would think but I can't quite figure it out. This answer Sort and group a list of dictionaries was pretty similar but not exactly what I need. It would be pretty simple if it were just one value but multiples are giving me fits. Any thoughts?

EDIT: So this hideous construction works. The thing is I have at least 13 fields (value1...value13) to do this on so need to make this more flxible.

list = [ {'id': 101, 'value1': 'yes', 'value2': '',    'value3': '',    'value4': 'no' },
         {'id': 102, 'value1': '',    'value2': 'yes', 'value3': '',    'value4': 'no' },
         {'id': 103, 'value1': 'no',  'value2': '',    'value3': 'yes', 'value4': ''   },
         {'id': 104, 'value1': 'yes', 'value2': '',    'value3': 'no',  'value4': 'yes'},
         {'id': 105, 'value1': '',    'value2': 'yes', 'value3': 'yes', 'value4': ''   },
         {'id': 106, 'value1': 'yes', 'value2': '',    'value3': '',    'value4': ''   }]

final_list = []
matched = False
for dict1 in list:
    sub_list = []
    for dict2 in list:
        if dict1 == dict2:
            continue
        print(dict1)
        print(dict2)
        print('---')
        if ((dict1['value1'] == dict2['value1'] or dict1['value1'] == '' or dict2['value1'] == '') and
            (dict1['value2'] == dict2['value2'] or dict1['value2'] == '' or dict2['value2'] == '') and
            (dict1['value3'] == dict2['value3'] or dict1['value3'] == '' or dict2['value3'] == '') and
            (dict1['value4'] == dict2['value4'] or dict1['value4'] == '' or dict2['value4'] == '')):
            # so these two match
            # now make sure it doesn't conflict with the other entries already there
            if sub_list:
                subsublist = sub_list
                sub_conflict = False
                for dict3 in subsublist:
                    if dict2 == dict3:
                        continue
                    print("    ",dict2)
                    print("    ",dict3)
                    if ((dict2['value1'] == dict3['value1'] or dict2['value1'] == '' or dict3['value1'] == '') and
                        (dict2['value2'] == dict3['value2'] or dict2['value2'] == '' or dict3['value2'] == '') and
                        (dict2['value3'] == dict3['value3'] or dict2['value3'] == '' or dict3['value3'] == '') and
                        (dict2['value4'] == dict3['value4'] or dict2['value4'] == '' or dict3['value4'] == '')):
                        print('no conflict for this one')
                    else:
                        sub_conflict = True
                if not sub_conflict:
                    sub_list.append(dict2)
            else:
                sub_list.append(dict1)
                sub_list.append(dict2)
                print('appending both to list')
    #if not matched and [dict1 not in list3 for list3 in final_list]:
    # if not matched:
    #     print('not matched')
    #     sub_list=[dict1]
    sorted_list = sorted(sub_list, key=lambda k: k['id'])
    print('-------------------------------')
    final_list.append(tuple(sorted_list))

#print(*final_list)

final_final_list = []
for list in final_list:
    if list not in final_final_list:
        final_final_list.append(list)

for list in final_final_list:
    # print(list)
    for list2 in list:
        print(list2)
    print('')

Solution

  • Assuming your list of dicts is called dict_list, does this

    from itertools import product
    
    values = ['yes', 'no']
    num_value_fields = 4
    keys = [f'value{i}' for i in range(1, num_value_fields + 1)]
    results = []
    for combo in product(values, repeat=num_value_fields):
        result = [d for d in dict_list if all(d[key] in {value, ''}
                                              for key, value in zip(keys, combo))]
        if len(result) > 1:
            results.append(result)
    

    produce your expected output

    [[{'id': 105, 'value1': '', 'value2': 'yes', 'value3': 'yes', 'value4': ''},
      {'id': 106, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': ''}],
     [{'id': 101, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': 'no'},
      {'id': 102, 'value1': '', 'value2': 'yes', 'value3': '', 'value4': 'no'},
      {'id': 105, 'value1': '', 'value2': 'yes', 'value3': 'yes', 'value4': ''},
      {'id': 106, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': ''}],
     [{'id': 104, 'value1': 'yes', 'value2': '', 'value3': 'no', 'value4': 'yes'},
      {'id': 106, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': ''}],
     [{'id': 101, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': 'no'},
      {'id': 102, 'value1': '', 'value2': 'yes', 'value3': '', 'value4': 'no'},
      {'id': 106, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': ''}],
     [{'id': 101, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': 'no'},
      {'id': 106, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': ''}],
     [{'id': 104, 'value1': 'yes', 'value2': '', 'value3': 'no', 'value4': 'yes'},
      {'id': 106, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': ''}],
     [{'id': 101, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': 'no'},
      {'id': 106, 'value1': 'yes', 'value2': '', 'value3': '', 'value4': ''}],
     [{'id': 103, 'value1': 'no', 'value2': '', 'value3': 'yes', 'value4': ''},
      {'id': 105, 'value1': '', 'value2': 'yes', 'value3': 'yes', 'value4': ''}],
     [{'id': 102, 'value1': '', 'value2': 'yes', 'value3': '', 'value4': 'no'},
      {'id': 103, 'value1': 'no', 'value2': '', 'value3': 'yes', 'value4': ''},
      {'id': 105, 'value1': '', 'value2': 'yes', 'value3': 'yes', 'value4': ''}]]
    

    ?