switching keys and values in a dictionary in python

Question

Say I have a dictionary like so:

my_dict = {2:3, 5:6, 8:9}

Is there a way that I can switch the keys and values to get:

{3:2, 6:5, 9:8}

What about `my_new_dict = dict(zip(my_dict.values(), my_dict.keys()))` — nauer, Sep 08 '20 at 20:52

score 125 · Accepted Answer · edited Apr 27 '22 at 14:40

125

For Python 3:

my_dict2 = {y: x for x, y in my_dict.items()}

For Python 2, you can use

my_dict2 = dict((y, x) for x, y in my_dict.iteritems())

edited Apr 27 '22 at 14:40

Nico Schlömer

53,797
27
201
249

answered Nov 29 '11 at 03:39

GWW

43,129
11
115
108

8

`.iteritems()` is not valid on Python 3.x... – JBernardo Nov 29 '11 at 03:47
@JBernardo: Fixed thanks I forgot about that – GWW Nov 29 '11 at 03:48
3

another correction - doing the dictionary comprehension you suggest `{(y, x) for x, y in dic_cols_w.items()}` returns a set, not a dictionary. You would still need to wrap this in a `dict(...)` to get a dictionary – tsando Apr 05 '17 at 13:06
4

@tsando: I think you have read my answer incorrectly. I have `{y:x ...` – GWW Apr 05 '17 at 22:34
How would you do this if the values were a list of items and each item in the list should be a key? – Woody Pride Feb 06 '19 at 14:49
I am not sure I understand. Perhaps ask a new question or look for another answer? – GWW Feb 06 '19 at 21:48
In Python 3.6, it (2nd solution) still gives an error: `AttributeError: 'dict' object has no attribute 'iteritems'`. Any workaround? Thank you in advance! – Milan Feb 09 '22 at 14:54
1

For the future readers: for Python 3.x, use `item()` instead of `iteritems()` i.e. `my_dict2 = {y:x for x,y in my_dict.items()}`. Source: https://stackoverflow.com/a/483833/10358768 – Milan Feb 09 '22 at 14:57

score 14 · Answer 2 · edited Jul 19 '16 at 11:39

14

my_dict = { my_dict[k]:k for k in my_dict}

edited Jul 19 '16 at 11:39

Ingo Karkat

167,457
16
250
324

answered Jul 19 '16 at 11:21

VladEg

141
1
2

3

Clean answer this one. – gosuto Feb 28 '20 at 18:16

score 9 · Answer 3 · answered Nov 29 '11 at 03:40

9

Try this:

my_dict = {2:3, 5:6, 8:9}

new_dict = {}
for k, v in my_dict.items():
    new_dict[v] = k

answered Nov 29 '11 at 03:40

Óscar López

232,561
37
312
386

score 7 · Answer 4 · edited May 23 '17 at 12:10

Use this code (trivially modified) from the accepted answer at Python reverse / invert a mapping:

dict((v,k) for k, v in my_dict.iteritems())

Note that this assumes that the values in the original dictionary are unique. Otherwise you'd end up with duplicate keys in the resulting dictionary, and that is not allowed.

And, as @wim points out, it also assumes the values are hashable. See the Python glossary if you're not sure what is and isn't hashable.

score 6 · Answer 5 · answered Nov 29 '11 at 03:41

6

maybe:

flipped_dict = dict(zip(my_dict.values(), my_dict.keys()))

answered Nov 29 '11 at 03:41

aaron

1,068
12
14

2

this is probably not safe as it relies on key and value ordering, although it happened to work for me. I like JBernardo/GWW's answer(s). – aaron Nov 29 '11 at 03:43
It's probably better to use `iterkeys` and `itervalues` – GWW Nov 29 '11 at 03:44
That's safe if the dict isn't changed during the operation. I liked because you did it without a comprehension. – JBernardo Nov 29 '11 at 03:51

score 3 · Answer 6 · answered Jan 19 '18 at 10:25

3

First of all it is not guaranteed that this is possible, since the values of a dictionary can be unhashable.

In case these are not, we can use a functional approach with:

reversed_dict = dict(map(reversed, original_dict.items()))

answered Jan 19 '18 at 10:25

Willem Van Onsem

443,496
30
428
555

score 2 · Answer 7 · answered Jul 29 '16 at 20:48

Sometimes, the condition that the values are all unique will not hold, in which case, the answers above will destroy any duplicate values.

The following rolls the values that might be duplicates up into a list:

from itertools import count
dict([(a,[list(d.keys())[i] for i,j in zip(count(), d.values())if j==a in set(d.values())])

I'm sure there's a better (non-list-comp) method, but I had a problem with the earlier answers, so thought I'd provide my solution in case others have a similar use-case.

P.S. Don't expect the dict to remain neatly arranged after any changes to the original! This method is a one-time use only on a static dict - you have been warned!

score 2 · Answer 8 · answered Dec 23 '18 at 03:56

2

If the values are not unique this will collide the key space in conversion. Best is to keep the keys in list when switching places

below handles this -

RvsD = dict()
for k,v in MyDict.iteritems():
    RsvD.setdefault(v, []).append(k)

answered Dec 23 '18 at 03:56

mungayree

313
1
12

score 1 · Answer 9 · edited Mar 03 '23 at 08:44

1

You can do it like this:

Function:

def inverse_dict(my_dict):
    updated_dict = {}
  
    for i in my_dict:
        updated_dict[my_dict[i]] = [i]

    return updated_dict

Main():

def main():
    year = {'day': 15, 'month': 3, 'year': 2019}
    print(inverse_dict(year))

if __name__ == "__main__":
    main()

Ouput:

{15: ['day'], 3: ['month'], 2019: ['year']}

edited Mar 03 '23 at 08:44

Masthan

727
1
8
29

answered Mar 15 '19 at 12:26

Alon Eilat

53
6

score 1 · Answer 10 · answered Jun 23 '19 at 11:55

If values are not unique in the dictionary, then you can map keys to that value. Here an example of how to do it using defaultdict

from collections import defaultdict


def reverse_dict(data: dict) -> dict:
    rd = defaultdict(list)
    for k, v in data.items():
        rd[v].append(k)
    return rd


if __name__ == "__main__":
    from collections import Counter

    data = "aaa bbb ccc ddd aaa bbb ccc aaa"
    c = Counter(data.split())
    print(c)
    # Counter({'aaa': 3, 'bbb': 2, 'ccc': 2, 'ddd': 1})
    rd = reverse_dict(c)
    print(rd)
    # defaultdict(<class 'list'>, {3: ['aaa'], 2: ['bbb', 'ccc'], 1: ['ddd']})

switching keys and values in a dictionary in python

10 Answers10

Linked

Related