Dict not hashable python

Question

I looked online and cant seem to understand much of it. im new to python and was wondering how i can fix this.

when running:

results = getRecommendations(userCompare[0], userCompare[0]['1'], sim_distance)

i get error:

    TypeError                                 Traceback (most recent call last)
<ipython-input-147-4d74cac55074> in <module>()
----> 1 results = getRecommendations(userCompare[0], userCompare[0]['1'], sim_distance)

<ipython-input-54-5f2d7e0dd3ba> in getRecommendations(data, person, similarity)
      5     for other in data:
      6         if other==person: continue #dont compare self
----> 7         sim=similarity(data, person, other)
      8         if sim<=0: continue #ignore scores of 0 or lower
      9         for item in data[other]:

<ipython-input-146-b30288308fee> in sim_distance(data, c1, c2)
      2 def sim_distance(data, c1, c2):
      3     si = {} #get the list of shared items
----> 4     for item in data[c1]:
      5         if item in data[c2]:
      6             si[item] = 1

TypeError: unhashable type: 'dict'

to create userCompare i did the following:

 movies = {}
    prefsList = []
    def loadMovieLens(path = directory):
        # Get movie titles
        for line in open(path + 'u.item'):
            (id, title) = line.split('|')[0:2]
            movies[id] = title 
        # Load data
        for k in range(len(centroidsM)):
            prefs ={}
            for rows in range(len(centroidsM[k])):
                for columns in range(len(centroidsM[k][0,:])):
                    user = str(rows+1)
                    movieid =str(columns+1)
                    prefs.setdefault(user,{})
                    prefs[user][movies[movieid]] = float(centroidsM[k][rows,columns])
            prefsList.append(prefs)
        return prefsList

I basically had an array of centroids with different K values, each K value has is a kx1682 matrix (k meaning number of clusters) so i loaded that into a list of dicts. i hope this makes sense. im starting to hate python or atleast dicts.

Thomas Orozco · Accepted Answer · 2013-11-21 08:24:33Z

1

You can't use a dict as a dictionary key. What would happen if I did:

d = {}
k1 = {1: 2}
k2 = {2: 1}

d[k1] = "a"
d[k2] = "b"

k1[2] = 1
k2[1] = 2

I now have k2 == k1, so what does d[{1:2, 2:1}] do? Well, that's why you can't use a dict as a key.

If you really need to do that (e.g. to use in a Counter), here's an option: freezing the dict:

#coding:utf-8
FROZEN_TAG = "__frozen__"


def freeze_dict(obj):
    if isinstance(obj, dict):
        dict_items = list(obj.items())
        dict_items.append((FROZEN_TAG, True))
        return tuple([(k, freeze_dict(v)) for k, v in dict_items])
    return obj


def unfreeze_dict(obj):
    if isinstance(obj, tuple):
        if (FROZEN_TAG, True) in obj:
            out = dict((k, unfreeze_dict(v)) for k, v in obj)
            del out[FROZEN_TAG]
            return out
    return obj

It's from here.

answered Nov 21, 2013 at 8:24

Thomas Orozco

55.6k12 gold badges120 silver badges120 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

user1819227 Over a year ago

Thank you! actually i figured out my problem... i was suppose to put in '1' or whatever variable in the 2nd argument when calling the function.. saved me alot of time and frustration

HeyWatchThis Over a year ago

That's a good alternative idea. An out-of-the-box way to make "hashable-dicts" is to just json.dump() your dict before using it as a key.

aIKid · Accepted Answer · 2013-11-21 08:31:30Z

1

You're getting that error because you are using an "unhashable" value (dict) as a dictionary key.

For example:

{dic():1}

Will give then same error.

Dictionaries cannot be used for keys, since they are mutable. Allowing mutable keys would wreak havoc with the dictionary, since it would not know when the keys had changed, so lookups would be impossible.

edited Nov 21, 2013 at 8:31

aIKid

28.5k5 gold badges41 silver badges65 bronze badges

answered Nov 21, 2013 at 8:24

Paul Draper

84.2k53 gold badges216 silver badges303 bronze badges

1 Comment

georg Over a year ago

Generally, mutability != hashability. It's fine to use a mutable object as a key as long as it provides a consistent hash function.

Collectives™ on Stack Overflow

Dict not hashable python

2 Answers 2

2 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related