在嵌套字典中通过求和排序python pandas

Question

我已经嵌套了一个像这样的字典

my_dict=  {'name1': {'code1': {'brand1': 2}},'name2': {'code2.1': {'brand2.1': 2,'brand2.2': 8,'brand2.3': 5, 'brand2.4': 4},'code2.2': {'brand2.1': 2, 'brand1': 1, 'brand2.5': 25}},'name3': {'code1': {'brand2.1': 2},'code3': {'brand4': 1,'brand3.1':2}}}    

我需要在"code"与依赖于价值总和的"品牌"。例如,

target_dict=  {'name1': {'code1': {'brand1': 2}}, 'name2': {'code2.2': {'brand2.1':2,'brand1': 1,'brand2.5': 25},'code2.1': {'brand2.1': 2,'brand2.2': 8,'brand2.3': 5,'brand2.4': 4}}, 'name3': {'code3': {'brand4': 1, 'brand3.1':2},'code1': {'brand2.1': 2}}}    

*# 'code2.2'优先，因为2+1+25=28>2 + 8 + 5 + 4 = 19# 'code3' first because 1+2=3>2

我可以加值"品牌"通过"code">

sum_values = [[[i, sum(v[i].values())] for i in v.keys()] for x,y in v.items() for k,v in my_dict.items()]

并尝试与sort函数组合为

target_dict = sorted(my_dict.items(), key=lambda i: [[[i, sum(v[i].values())] for i in v.keys()] for x,y in v.items() for k,v in my_dict.items()], reverse=True).

感谢您的关注和帮助!

Answer 1

尝试(假设Python的版本足以保持dict的创建顺序):

my_dict = {
"name1": {"code1": {"brand1": 2}},
"name2": {
"code2.1": {"brand2.1": 2, "brand2.2": 8, "brand2.3": 5, "brand2.4": 4},
"code2.2": {"brand2.1": 2, "brand1": 1, "brand2.5": 25},
},
"name3": {"code1": {"brand2.1": 2}, "code3": {"brand4": 1, "brand3.1": 2}},
}
out = {
k: dict(sorted(v.items(), key=lambda d: sum(d[1].values()), reverse=True))
for k, v in my_dict.items()
}
print(out)

打印:

{
"name1": {"code1": {"brand1": 2}},
"name2": {
"code2.2": {"brand2.1": 2, "brand1": 1, "brand2.5": 25},
"code2.1": {"brand2.1": 2, "brand2.2": 8, "brand2.3": 5, "brand2.4": 4},
},
"name3": {"code3": {"brand4": 1, "brand3.1": 2}, "code1": {"brand2.1": 2}},
}