How to return dictionary keys as a list in Python 3.3

http://btmiller.com/2015/04/13/get-list-of-keys-from-dictionary-in-python-2-and-3.html

Get a List of Keys From a Dictionary in Both Python 2 and Python 3

It was mentioned in an earlier post that there is a difference in how the keys() operation behaves between Python 2 and Python 3. If you’re adapting your Python 2 code to Python 3 (which you should), it will throw a TypeError when you try to operate on keys() like a list. So, if you depend on getting a list returned from keys(), here’s how to make it work for both Python 2 and Python 3.

In Python 2, simply calling keys() on a dictionary object will return what you expect:

$ python
>>> foo = { 'bar': "hello", 'baz': "world" }
>>> type(foo.keys())
<type 'list'>
>>> foo.keys()
['baz', 'bar']
>>> foo.keys()[0]
'baz'

That’s great, however, in Python 3, keys() no longer returns a list, but a view object:

The objects returned by dict.keys(), dict.values() and dict.items() are view objects. They provide a dynamic view on the dictionary’s entries, which means that when the dictionary changes, the view reflects these changes.

The dict_keys object is an iterator and looks a lot more like a set than a list. So using the same call in Python 3 would produce this result:

$ python3
>>> foo = { 'bar': "hello", 'baz': "world" }
>>> type(foo.keys())
<class 'dict_keys'>
>>> foo.keys()
dict_keys(['baz', 'bar'])
>>> foo.keys()[0]
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
TypeError: 'dict_keys' object does not support indexing

The TypeError can be avoided and compatibility can be maintained by simply converting the dict_keys object into a list which can then be indexed as normal in both Python 2 and Python 3:

$ python3
>>> foo = { 'bar': "hello", 'baz': "world" }
>>> type(list(foo.keys()))
<class 'list'>
>>> list(foo.keys())
['baz', 'bar']
>>> list(foo.keys())[0]
'baz'

And just for good measure, here it is in Python 2:

$ python
>>> foo = { 'bar': "hello", 'baz': "world" }
>>> type(list(foo.keys()))
<class 'list'>
>>> list(foo.keys())
['baz', 'bar']
>>> list(foo.keys())[0]
'baz'

http://stackoverflow.com/questions/16819222/how-to-return-dictionary-keys-as-a-list-in-python-3-3

65down votefavorite

6

I noticed something very weird - or let's say, something that is very different from Python 2.7 and older versions of Python 3 I believe.

Previously, I could get dictionary keys, values, or items of a dictionary very easily as list:

PYTHON 2.7
>>> newdict = {1:0, 2:0, 3:0}
>>> newdict
{1: 0, 2: 0, 3: 0}
>>> newdict.keys()
[1, 2, 3]

Now, I get something like this in

PYTHON 3.3.0
>>> newdict.keys()
dict_keys([1, 2, 3])

I am wondering if there is a way to return a list as I showed it in the Python 2.7 example. Because now, I have to do something like

newlist = list()
for i in newdict.keys():
    newlist.append(i)

EDIT:

Thanks, list(newdict.keys()) works as I wanted!

But there is another thing that bugs me now: I want to create a list of reversed dictionary keys and values to sort them by values. Like so (okay, this is a bad example, because the values are all 0 here)

>>> zip(newdict.values(), newdict.keys())
[(0, 1), (0, 2), (0, 3)]

However, in Python3 I get something like

>>> zip(list(newdict.keys()), list(newdict.values()))
<zip object at 0x7f367c7df488>

Okay, sorry, I just figured out that you have to use a list() function around zip() too.

list(zip(newdict.values(), newdict.keys()))
[(0, 1), (0, 2), (0, 3)]

This is really something one has to get used to

list dictionary python-3.x python-2.x

shareimprove this question

edited May 29 '13 at 16:38

asked May 29 '13 at 16:24   user2015601

1

If you're trying to sort a dictionary by values, try this oneliner: sorted(newdict.item‌​s(),key=lambda x: x[1]). newdict.items() returns the key-value pairs as tuples (just like you're doing with the zip above).sorted is the built-in sort function and it permits a key parameter which should be a function that transforms each list element into the value which should be used to sort. – ChrisMay 29 '13 at 17:33 

Looks very handy, thanks! – user2015601 May 29 '13 at 18:54

Interesting thread safety issue regarding this topic is here: blog.labix.org/2008/06/27/… – Paul May 10 at 18:00

add a comment

3 Answers

activeoldestvotes

up vote78down voteaccepted

Try list(newdict.keys()).

This wil convert the dict_keys object to a list.

On the other hand, you should ask yourself whether or not it matters. The Pythonic way to code is to assume duck typing (if it looks like a duck and it quacks like a duck, it's a duck). the dict_keys object will act like a list for most purposes. For instance:

for key in newdict.keys():
  print(key)

Obviously insertion operators may not work, but that doesn't make much sense for a list of dictionary keys anyway.

shareimprove this answer

answered May 29 '13 at 16:25 Chris 1,787714

Thank you for the quick response, it works! Regarding the second part of your answer: I think it matters for what I want to do with the list(s), I updated my question under the EDIT section. Thanks! – user2015601 May 29 '13 at 16:31 

1

newdict.keys() does not support indexing – Miguel de Val-Borro Sep 10 '14 at 17:54

5

list(newdict) also works (at least in python 3.4). Is there any reason to use the .keys() method? – naught101 Mar 31 '15 at 11:58