Changing the value in one key changes the values in all keys in a dictionary of sets created using fromkeys()

Question

I am initializing a sequence like this

seq = {'a', 'b', 'c', 'd', 'e'}

Now I am using fromkeys() to convert sequence to dict of sets. here is what I am doing:

val = set()
seq_dict =  dict.fromkeys(seq, val)

Now it seems if add an element to only one of my dictionary keys set, that element is being added to all the other sets. Here is the example:

seq_dict['a'].add("val1")
print(seq_dict)
{'d': {'val1'}, 'c': {'val1'}, 'b': {'val1'}, 'a': {'val1'}, 'e': {'val1'}}

Not sure if I am using the fromkeys in a right way though?

Answer 1

The fromkeys method is working as expected, since you have the correct result: a dictionary with keys that are the characters in seq variable, and each value that is an empty set.

Your issue is due to mutability of the value you assigned: you assigned an empty set as value for all keywords. If you change this (seq_dict['a'].add("val1")), then the values for all keywords are changed (each keyword references the original empty set).

You may want to use dictionary comprehension (which creates a new different copy of the empty set for each key) to avoid this issue:

seq = {'a', 'b', 'c', 'd', 'e'}
seq_dict = {key: set() for key in seq}

Answer 2

val = set()
seq_dict =  dict.fromkeys(seq, val)

You are initializing the seq_dict with the same set (a mutable type), so editing seq_dict['a'] you are editing the same set which is the value of all the other keys.

---

If you don't want this to happen, you'll have to initialize the dictionary in an other way.

>>> d = {i: set() for i in ('a', 'b', 'c', 'd', 'e')}
>>> d
{'a': set(), 'b': set(), 'c': set(), 'd': set(), 'e': set()}
>>> d['a']
set()
>>> d['a'].add('ehi')
>>> d
{'a': {'ehi'}, 'b': set(), 'c': set(), 'd': set(), 'e': set()}

---

So, in other words, you have to "isolate" the **immutable**s parts of the set:

seq = {'a', 'b', 'c', 'd', 'e'}
seq_dict = {key: set() for key in seq}