我有一个OrderedDict,我需要在保持排序的同时添加一个元素
import sys
import bisect
from collections import OrderedDict
arr = {('a',1111),('b',2222),('f',3333)}
arr = OrderedDict(arr)
bisect.insort(arr,('c',4444))
#expectedly arr = {('a',1111),('b',2222),('c',4444),('f',3333)}
#but actually TypeError: collections.OrderedDict is not a sequence
更新:我需要按键排序存储项目 但与
import sys
import bisect
from collections import OrderedDict
from sortedcontainers import sorteddict
arr = {('a',1111),('b',2222),('f',3333)}
arr = OrderedDict(arr)
arr.update({'c':4444}) #or arr['c'] = 4444
print(arr)
OrderedDict([('b', 2222), ('f', 3333), ('a', 1111), ('c', 4444)])
改为 OrderedDictх([('a',1111),('b',2222),('c',4444),('f',3333)])
像c++中的 map
最佳答案
将新项添加到原始项中,排序,生成新字典:
>>> arr = {('a',1111),('b',2222),('f',3333)}
>>> arr = collections.OrderedDict(arr)
>>> new = ('c',4444)
>>> items = list(arr.items())
>>> items.append(new)
>>> items.sort()
>>> arr = collections.OrderedDict(items)
>>> arr
OrderedDict([('a', 1111), ('b', 2222), ('c', 4444), ('f', 3333)])
或者更复杂的选项:
- 子类 collections.OrderedDict
- 使用
move_to_end
方法作为指导,创建一个将遍历双向链表的新方法;找到插入的地方;然后插入新键——也许可以在这里使用 bisect 或其他一些双向链表排序插入算法 - 覆盖
__setitem__
方法并在中调用新方法 - 或者只是替换 add-new-key-to-the-end 使用您在上一项目符号中提出的算法编写代码。
维护键排序顺序的排序字典
我无法弄清楚如何使 OrderedDict 子类工作 - 它有许多属性名称被破坏 - 只需要覆盖一两个方法,我不想花时间弄清楚名称修改方面。
因此只需从源代码中复制整个 OrderedDict 类from here - to here到一个单独的模块中,因此您可以导入它,并包含这些导入。
from _weakref import proxy as _proxy
from collections import _Link, _OrderedDictKeysView
from collections import _OrderedDictItemsView, _OrderedDictValuesView
import _collections_abc
from _weakref import proxy as _proxy
from reprlib import recursive_repr as _recursive_repr
from operator import itemgetter as _itemgetter, eq as _eq
import bisect
然后在类中更改以下内容:
- 类(class)名称当然可以随心所欲。类名后面是文档字符串和一些描述类行为的注释 - 这些应该更新。
class SortOrderedDict(dict):
- 覆盖
__setitem__
方法。下面使用bisect
来查找插入顺序。不知道是否真的有必要,它必须先列出字典键 View ,但那部分应该是快速的 C 代码(?在这里猜测)
def __setitem__(self, key, value,
dict_setitem=dict.__setitem__, proxy=_proxy, Link=_Link):
'od.__setitem__(i, y) <==> od[i]=y'
# Setting a new item creates a new link in the linked list,
# inserted at its key sorted position - uses less than comparisons,
# and the inherited dictionary is updated with the new key/value pair.
if key not in self:
self.__map[key] = link = Link()
root = self.__root
last = root.prev
link.key = key
curr = root.next
if curr is root: # first item!
link.prev, link.next = last, root
last.next = link
root.prev = proxy(link)
elif link.key < root.next.key: # at the beginning?
#print(f'{link.key} before {root.next.key}')
soft_link = root.next
link.prev, link.next = root, soft_link
soft_link.prev = link
root.next = link
elif root.prev.key < link.key: # at the end?
#print(f'{link.key} at the end after {root.prev.key}')
soft_link = root.prev
link.prev, link.next = soft_link, root
soft_link.next = link
root.prev = proxy(link)
else: # in the middle somewhere - use bisect
keys = list(self.keys())
i = bisect.bisect_left(keys,key)
right = self.__map[keys[i]]
#print(f'{link.key} between {right.prev.key} and {right.key}')
soft_link = right.prev
link.prev,link.next = soft_link,right
right.prev = link
soft_link.next = link
dict_setitem(self, key, value)
- 添加
update
方法 - 此类是dict
的子类,它会覆盖其更新方法,迫使它使用__setitem__
。
def update(self,other):
try:
other = other.items()
except AttributeError:
pass
for k,v in other:
self[k] = v
- 将此行
update = __update = _collections_abc.MutableMapping.update
更改为
__update = update
- 在
__reduce__
方法中,将for k in vars(OrderedDict()):
中的类名更改为您命名类的任何名称
for k in vars(SortOrderedDict()):
__eq__
方法中的相同内容。将if isinstance(other, OrderedDict):
更改为
if isinstance(other, SortOrderedDict):
如果使用 bisect 似乎不值得,只需遍历链表直到找到插入点。 (上面列出的所有其他更改仍然适用)
def __setitem__(self, key, value,
dict_setitem=dict.__setitem__, proxy=_proxy, Link=_Link):
'od.__setitem__(i, y) <==> od[i]=y'
# Setting a new item creates a new link in the linked list,
# inserted at its key sorted position - uses less than comparisons,
# and the inherited dictionary is updated with the new key/value pair.
if key not in self:
self.__map[key] = link = Link()
root = self.__root
last = root.prev
link.key = key
curr = root.next
if curr is root: # first item!
link.prev, link.next = last, root
last.next = link
root.prev = proxy(link)
# traverse the linked list; find sorted insertion point; insert
while curr is not root:
if link.key < curr.key:
soft_link = curr.prev
soft_link.next = link
link.prev = soft_link
link.next = curr
curr.prev = link
break
elif curr.next is root:
link.prev, link.next = curr, root
curr.next = link
root.prev = proxy(link)
break
curr = curr.next
dict_setitem(self, key, value)
用法
>>> arr = {('a',1111),('f',3333),('b',2222)}
>>> arr = SortOrderedDict(arr)
>>> arr
SortOrderedDict([('a', 1111), ('b', 2222), ('f', 3333)])
>>> other = {k:v for k,v in zip('tvsnpqkl',range(8))}
>>> arr.update(other)
>>> arr
SortOrderedDict([('a', 1111), ('b', 2222), ('f', 3333), ('k', 6), ('l', 7), ('n', 3), ('p', 4), ('q', 5), ('s', 2), ('t', 0), ('v', 1)])
>>> b = SortOrderedDict((('a',1111),('f',3333),('b',2222)))
>>> b.update(other)
>>> arr == b
True
>>> b == arr
True
>>>
关于python - 如何将项目添加到 OrderedDict,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/61369485/