问题
I did a loop over multiple files to get as a result an OrderedDict for each one (possibility to have different keys, which means not the same from an OrderedDict to another). I want to write/ pack the result into the same pandas dataframe. So we will have all the different keys as column names, and each row of the dataframe will represent an OrderedDict.
I have already my results looking like this :
OrderedDict([('mrz_type', 'ID'), ('valid_score', 70), ('valid_composite', False), ('type', 'ID'), ('country', ''), ('number', ''), ('date_of_birth', '840927'), ('sex', 'F'), ('nom', ), ('prenom', ''), ('dep', ''), ('service', '1'), ('office', '056'), ('check_number', '7'), ('check_date_of_birth', '4'), ('check_composite', '9'), ('valid_number', True), ('valid_date_of_birth', True)])
OrderedDict([('mrz_type', 'PASP'), ('valid_score', 62), ('valid_composite', False), ('type', 'P'), ('country', ''), ('number', ''), ('date_of_birth', '550912'), ('expiration_date', '200801'), ('nationality', ''), ('sex', 'M'), ('nom', ''), ('prenom', ''), ('check_number', '2'), ('check_date_of_birth', '9'), ('check_expiration_date', '1'), ('check_composite', '8'), ('valid_number', True), ('valid_date_of_birth', False), ('valid_expiration_date', True)])
OrderedDict([('mrz_type', 'IR'), ('valid_score', 28), ('valid_composite', False), ('type', 'IR'), ('country', ''), ('number', ''), ('date_of_birth', '750612'), ('expiration_date', '010119'), ('nationality', ''), ('sex', 'Z'), ('nom', ''), ('prenom', ''), ('num_etrg', ''), ('check_number', '6'), ('check_date_of_birth', '1'), ('check_expiration_date', ''), ('check_composite', ''), ('valid_number', False), ('valid_date_of_birth', True), ('valid_expiration_date', False)])
回答1:
following your the 3 examples of OrderedDicts you provided Note that the first OrderedDict you provided has a single element tuple ('nom', ) which I changed to ('nom',' ').
following 4 steps to achieve desired result:
list_of_ordered_dicts = [od1,od2,od3]
# flat all items for dicts together in a list
all_items =[item for dict_items in list_of_ordered_dicts
for item in dict_items.items() ]
# create a set with all columns
all_columns = set(dict(all_items).keys())
# update ordered dicts with new columns, setting missing columns values to None
for ordered_dict in list_of_ordered_dicts:
missing_columns = all_columns - set(ordered_dict.keys())
for column in missing_columns:
ordered_dict.setdefault(column, None)
# create dataframe
df = pd.DataFrame(list_of_ordered_dicts, columns=all_columns)
来源：https://stackoverflow.com/questions/57061143/how-to-pack-multilines-results-ordereddict-with-differents-keys-into-pandas-da