合并两个字典后，形成结果字典的子项是无序的。子项是月份 ['jan'， 'feb'， 'mar'， ...， 'dec']。在某些情况下，原始字典可能不包含子项(月(，因此输出会随着合并而消失。
所以我有两个字典，都有下一个结构{Model:{'Jan':[1], 'Feb':[2], Jun: [5], ...}
正如您在此示例中所看到的，某些子项(月(未表示，因此无法在原始字典中找到它们。但是，我需要的是合并的字典来保持每月订单，无论原始字典的外观如何。
合并功能：
def merge_dicts(dict1, dict2):
'''Marge two dicts by adding up (accumulating) values in each key.
Returns: A merge (addition) of two dictionaries by adding values of same keys
'''
# Merge dictionaries and add values of same keys
out = {**dict1, **dict2}
for key, value in out.items():
if key in dict1 and key in dict2:
out[key] = [value, dict1[key]]
#Things got harder, out[key] appends in list of list of list... no itertools can help here.
lst = [] #the one dimensional list to fix the problem of list of list with out[key]
for el in out[key]:
try:
#if inside out[key] there is a list of list we split it
for sub_el in el:
lst.append(sub_el)
except:
#if inside out[key] there is only a single float
lst.append(el)
#Replace the old key with the one dimensional list
out[key] = lst
return out
我如何合并它：
for c in range(len([*CMs_selection.keys()])):
if c == 0:
#First merge, with dict0 & dict1
merged_dict = {cm:merge_dicts(CMs_selection[[*CMs_selection.keys()][c]][cm], CMs_selection[[*CMs_selection.keys()][c + 1]][cm])
for cm in CMs_selection[[*CMs_selection.keys()][0]]}
elif c > 0 and c < (len(years) - 1):
#Second merge to n merge, starting with dict_merged and dict 2
merged_dict = {cm:merge_dicts(merged_dict[cm], CMs_selection[[*CMs_selection.keys()][c + 1]][cm])
for cm in CMs_selection[[*CMs_selection.keys()][0]]}
现在，在尝试了所有可能的合并之后，我总是得到这个结果。
{'Model1': {'Jan': [-0.0952586755156517,
0.1015196293592453,
-0.10572463274002075],
'Oct': [-0.02473766915500164,
0.0678798109292984,
0.08870666474103928,
-0.06378963589668274],
'Nov': [-0.08730728179216385,
0.013518977910280228,
0.023245899006724358,
-0.03917887806892395],
'Jul': [-0.07940272241830826, -0.04912888631224632, -0.07454635202884674],
'Dec': [-0.061335086822509766, -0.0033914903178811073, 0.09630533307790756],
'Mar': [0.029064208269119263, 0.11327305436134338, 0.009556809440255165],
'Apr': [-0.04433680325746536, -0.08620205521583557],
'Jun': [-0.036688946187496185, 0.05543896555900574, -0.07162825018167496],
'Aug': -0.03712410107254982,
'Sep': [0.007421047426760197, 0.008665643632411957],
'Feb': [-0.02879650704562664, 0.013025006279349327]},
'Model2': {'Feb': -0.05173473060131073,
'Jun': [-0.09126871824264526,
-0.09009774029254913,
0.10458160936832428,
-0.09445420652627945,
-0.04294373467564583],
'Aug': [-0.07917020469903946, 0.011026041582226753],
'Oct': [-0.10164830088615417, ....
....
随着无序的月份。请帮帮我！
如果我们只专注于合并字典，首先，我们需要定义月份的正常顺序，然后按该顺序进行合并，因为 Python 不知道这个顺序。如果第一个字典中不存在，则它不能在"Feb"和"Apr"之间添加"Mar"。所以，我们需要自己定义顺序。
此外，您需要两种不同的解决方案来合并浮点值和合并列表。我mode参数添加到我的解决方案中。
def merge_dicts(list_of_dicts, mode):
keys = set(key for d in list_of_dicts for key in d.keys())
months = ["Jan","Feb","Mar","Apr","May","Jun","Jul","Aug","Sep","Oct","Nov","Dec"]
ordered_keys = [month for month in months if month in keys]
out = {}
for key in ordered_keys:
out[key] = []
for d in list_of_dicts:
if key in d:
if mode == "append":
out[key].append(d[key])
elif mode == "extend":
out[key].extend(d[key])
return out
CMs_selection = {2006: {'Model1': {'Jan': -0.1, 'Oct': -0.063}, 'Model2': {'Feb': -0.051, 'Jun': -0.04, 'Oct': 0.07}, 'Model3': {'Mar': -0.030, 'Jun': 0.02, 'Aug': 0.0561,}, 'Model4': {'Feb': -0.026, 'Dec': -0.06}}, 2007: {'Model1': {'Jul': -0.07, 'Oct': 0.8, 'Nov': 0.38, 'Dec': 0.1}, 'Model2': {'Jun': -0.09, 'Aug': -0.079, 'Sep': -0.7}}}
for key in CMs_selection:
CMs_selection[key] = merge_dicts(CMs_selection[key].values(), "append")
print(CMs_selection)
result = merge_dicts(CMs_selection.values(), "extend")
print(result)
输出：
{2006: {'Jan': [-0.1], 'Feb': [-0.051, -0.026], 'Mar': [-0.03], 'Jun': [-0.04, 0.02], 'Aug': [0.0561], 'Oct': [-0.063, 0.07], 'Dec': [-0.06]}, 2007: {'Jun': [-0.09], 'Jul': [-0.07], 'Aug': [-0.079], 'Sep': [-0.7], 'Oct': [0.8], 'Nov': [0.38], 'Dec': [0.1]}}
{'Jan': [-0.1], 'Feb': [-0.051, -0.026], 'Mar': [-0.03], 'Jun': [-0.04, 0.02, -0.09], 'Jul': [-0.07], 'Aug': [0.0561, -0.079], 'Sep': [-0.7], 'Oct': [-0.063, 0.07, 0.8], 'Nov': [0.38], 'Dec': [-0.06, 0.1]}