获取嵌套字典与列表的组合数
Posted
技术标签:
【中文标题】获取嵌套字典与列表的组合数【英文标题】:Get number of combinations of nested dictionary with lists 【发布时间】:2021-01-13 13:49:52 【问题描述】:我有一个嵌套字典列表,里面有列表,我想通过所有字典计算所有可能组合的数量。
输入示例
json =
[
"lang": "fr",
"dates": ["d1", "d2"],
"address": [
"city": "Paris",
"zip": "75001"
,
"city": "Lyon",
"zip": "69600"
]
,
"lang": "fr",
"dates": ["d2", "d3"],
"address": [
"city": "Paris",
"zip": "75001"
]
]
预期的输出是:
'address.city_dates': "Paris_d1": 1, "Paris_d2": 2 ,"Paris_d3": 1, "Lyon_d1": 1, "Lyon_d2": 1,
'address.city_lang': "Paris_fr": 2, "Lyon_fr": 1,
'address.city_address.zip': "Lyon_69600": 1, "Paris_75001": 2,
'address.zip_dates': "75001_d1": 1, "75001_d2": 2 ,"75001_d3": 1, "69600_d1": 1, "69600_d2": 1,
'address.zip_lang': "75001_fr": 2, "69600_fr": 1,
'dates_lang': "d1_fr": 1, "d2_fr": ,"d3_fr": 1
注意:在我的例子中,关系是可以互换的:X_Y 类似于 Y_X
尝试: 我尝试了以下将被递归调用的函数,但我不知道如何处理列表并在所有级别上查找组合。
def get_combs(element, contextKey, relations):
if type(element) is list:
# for each element of the list, treat element
for tokenElement in element:
get_combs(tokenElement, contextKey, relations)
elif type(element) is dict:
keys = list(element.keys())
keys.sort()
for first_key in keys:
# remove current first key to avoid X_Y and Y_X being different
keys.remove(first_key)
keys.sort()
for second_key in keys:
key = first_key + "_" + second_key
value_key = str(element[first_key]) + "_" + str(element[second_key])
if not key in relations:
relations[key] = value_key: 1
else:
if not value_key in relations[key]:
relations[key][value_key] = 1
else:
relations[key][value_key] += 1
# recall function to deal wit elements inside the dict
get_combs(element[first_key], first_key, relations)
def main(json):
relations =
# loop over all the elements inside json
for element in json:
get_combs(element, 'root', relations)
return relations
pippo = main(json)
print(pippo)
输出:
'address_dates': "['city': 'Paris', 'zip': '75001', 'city': 'Lyon', 'zip': '69600']_['d1', 'd2']": 1,
"['city': 'Paris', 'zip': '75001']_['d2', 'd3']": 1,
'address_lang': "['city': 'Paris', 'zip': '75001', 'city': 'Lyon', 'zip': '69600']_fr": 1,
"['city': 'Paris', 'zip': '75001']_fr": 1,
'city_zip': "Lyon_69600": 1, "Paris_75001": 2,
'lang_dates': "fr_['d1', 'd2']": 1, "fr_['d2', 'd3']": 1
【问题讨论】:
你尝试过什么,你在哪里卡住了? 我刚刚用我尝试过的内容编辑了帖子 你的例子和你的预期结果不匹配,给一个更好的,我可以帮忙,但是看到你的 json 例子你有 Paris_d1,Paris_d2,Paris_d3 和 Paris_d4,Lyon_d1 和 Lyon_d2 都显示只有1个;您的预期结果不同:"Paris_d1": 1, "Paris_d2": 2 ,"Paris_d3": 1, "Lyon_d1": 1, "Lyon_d2": 1。请举例说明您的正确预期结果是什么 很抱歉复制粘贴不一致。我编辑了数据 您的输出中是否需要 lang_dates_zip ?你的 recurvive 函数对我来说真的很难,但仍然不明白你的问题/期望的输出:'想要通过所有 dicts 计算所有可能组合的数量。一开始 【参考方案1】:您可以递归地展平结构,然后按形成的键进行分组。从那里,可以应用基本组合:
from collections import defaultdict, Counter
import itertools as it
json = ['lang': 'fr', 'dates': ['d1', 'd2'], 'address': ['city': 'Paris', 'zip': '75001', 'city': 'Lyon', 'zip': '69600'], 'lang': 'fr', 'dates': ['d2', 'd3'], 'address': ['city': 'Paris', 'zip': '75001']]
def get_keys(d, c = [], j = None):
if not isinstance(d, (dict, list)):
yield ('.'.join(c), d, j)
elif isinstance(d, list):
yield from [i for j, b in enumerate(d) for i in get_keys(b, c = c, j = j if isinstance(b, dict) else None)]
else:
for a, b in d.items():
yield from get_keys(b, c = c+[a], j = j)
def get_combos(data):
d = defaultdict(list)
for a, *b in get_keys(data):
d[a].append(b)
return f'b_a':Counter([f'y_x' for [x, l1], [y, l2] in it.product(d[a], d[b]) if type(l1) != type(l2) or l1 == l2]) for a, b in it.combinations(d, 2)
new_d =
for i in json:
for a, b in get_combos(i).items():
for c, d in b.items():
new_d[a] = **(l:=new_d.get(a, )), c:d if c not in l else l[c]+d
输出:
'dates_lang': 'd1_fr': 1, 'd2_fr': 2, 'd3_fr': 1, 'address.city_lang': 'Paris_fr': 2, 'Lyon_fr': 1, 'address.zip_lang': '75001_fr': 2, '69600_fr': 1, 'address.city_dates': 'Paris_d1': 1, 'Lyon_d1': 1, 'Paris_d2': 2, 'Lyon_d2': 1, 'Paris_d3': 1, 'address.zip_dates': '75001_d1': 1, '69600_d1': 1, '75001_d2': 2, '69600_d2': 1, '75001_d3': 1, 'address.zip_address.city': '75001_Paris': 2, '69600_Lyon': 1
编辑:要打印扁平结构,首先存储 get_keys
调用:
def get_combos(data):
d = defaultdict(list)
t_result = list(get_keys(data))
print(t_result) #displaying flattened input dict
for a, *b in t_result:
d[a].append(b)
return f'b_a':Counter([f'y_x' for [x, l1], [y, l2] in it.product(d[a], d[b]) if type(l1) != type(l2) or l1 == l2]) for a, b in it.combinations(d, 2)
【讨论】:
抱歉打扰我是一个新手,你有一个易于理解的“扁平结构”的定义/解释,或者指出我可以在你的代码中打印扁平结构的位置。谢谢? 您的回答有效 谢谢!但是,您能否为变量提供重要的名称? (a, b, c..) 很难理解你的算法以上是关于获取嵌套字典与列表的组合数的主要内容,如果未能解决你的问题,请参考以下文章