如何按照插入顺序从字典中检索项目？

Question

是否可以按照插入的顺序从 Python 字典中检索项目？

Answer 1

The standard Python dict does this by default if you're using CPython 3.6+ (or Python 3.7+ for any other implementation of Python).

On older versions of Python you can use collections.OrderedDict .

Answer 2

Use OrderedDict(), available since version 2.7

Just a matter of curiosity:

from collections import OrderedDict
a = {}
b = OrderedDict()
c = OrderedDict()

a['key1'] = 'value1'
a['key2'] = 'value2'

b['key1'] = 'value1'
b['key2'] = 'value2'

c['key2'] = 'value2'
c['key1'] = 'value1'

print a == b  # True
print a == c  # True
print b == c  # False

Answer 3

As of Python 3.7, the standard dict preserves insertion order. From the docs :

Changed in version 3.7: Dictionary order is guaranteed to be insertion order. This behavior was implementation detail of CPython from 3.6.

So, you should be able to iterate over the dictionary normally or use popitem() .

Answer 4

The other answers are correct; it's not possible, but you could write this yourself. However, in case you're unsure how to actually implement something like this, here's a complete and working implementation that subclasses dict which I've just written and tested. (Note that the order of values passed to the constructor is undefined but will come before values passed later, and you could always just not allow ordered dicts to be initialized with values.)

class ordered_dict(dict):
    def __init__(self, *args, **kwargs):
        dict.__init__(self, *args, **kwargs)
        self._order = self.keys()

    def __setitem__(self, key, value):
        dict.__setitem__(self, key, value)
        if key in self._order:
            self._order.remove(key)
        self._order.append(key)

    def __delitem__(self, key):
        dict.__delitem__(self, key)
        self._order.remove(key)

    def order(self):
        return self._order[:]

    def ordered_items(self):
        return [(key,self[key]) for key in self._order]


od = ordered_dict()
od["hello"] = "world"
od["goodbye"] = "cruel world"
print od.order()            # prints ['hello', 'goodbye']

del od["hello"]
od["monty"] = "python"
print od.order()            # prints ['goodbye', 'monty']

od["hello"] = "kitty"
print od.order()            # prints ['goodbye', 'monty', 'hello']

print od.ordered_items()
# prints [('goodbye','cruel world'), ('monty','python'), ('hello','kitty')]

Answer 5

You can't do this with the base dict class -- it's ordered by hash. You could build your own dictionary that is really a list of key,value pairs or somesuch, which would be ordered.

Answer 6

Or, just make the key a tuple with time.now() as the first field in the tuple.

Then you can retrieve the keys with dictname.keys(), sort, and voila!

Gerry

Answer 7

I've used StableDict before with good success.

http://pypi.python.org/pypi/StableDict/0.2

Answer 8

Or use any of the implementations for the PEP-372 described here , like the odict module from the pythonutils .

I successfully used the pocoo.org implementation, it is as easy as replacing your

my_dict={}
my_dict["foo"]="bar"

with

my_dict=odict.odict()
my_dict["foo"]="bar"

and require just this file

Answer 9

除非您将密钥存储在单独的列表中以供以后参考，否则这是不可能的。

Answer 10

What you can do is insert the values with a key representing the order inputted, and then call sorted() on the items.

>>> obj = {}
>>> obj[1] = 'Bob'
>>> obj[2] = 'Sally'
>>> obj[3] = 'Joe'
>>> for k, v in sorted(obj.items()):
...     print v
... 
Bob
Sally
Joe
>>> 

Answer 11

如果您不需要 dict 功能，而只需要按照插入的顺序返回元组，那么队列不是更好用吗？