Appending to one python dictionary key appends to all for some reason

Question

This is my python code (I am using Python 2.7.6):

row_of_words = ['word 1', 'word 2']
hw = ['h1', 'h2']
TD = {}

TD = TD.fromkeys(hw, [])
# TD is now {'h1': [], 'h2': []}

for index in range(len(row_of_words)):
    # This loops twice and in the first loop, index = 0 and in the second
    # loop, index = 1

    # Append 1 to the value of 'h1' (which is a list [])
    # The value for 'h1' should now be [1].

    # This loops twice to append 1 twice to the value of 'h1'.
    TD['h1'].append(1)

>>>TD
{'h2': [1, 1], 'h1': [1, 1]}

As you can see in the two lines right above, when I print TD, it shows that the 1 was appended to both h1 and h2. Any idea why it was appended to both even though I mentioned to only append it to TD['h1']? How do I make it so that it only appends to the value of 'h1'? I commented my steps showing how I think it works.

Using your code I got the following error: `TypeError: descriptor 'append' requires a 'list' object but received a 'int'` — albert, Dec 01 '15 at 00:06
`dict.fromkeys(seq,value)` will set `value` to every key if `value` is provided which means the value of each key will be the exact same list if you pass in `value= []` — R Nar, Dec 01 '15 at 00:07
@user2719875: Using Python 3.4.3 - did not see your edit. Sorry for that. — albert, Dec 01 '15 at 00:09
@RNar oh, is there a way for me to non-manually set each key in the dictionary to a distinct empty list? — SilentDev, Dec 01 '15 at 00:11
Possible duplicate of [Python list of lists, changes reflected across sublists unexpectedly](http://stackoverflow.com/questions/240178/python-list-of-lists-changes-reflected-across-sublists-unexpectedly) — TessellatingHeckler, Dec 01 '15 at 00:15

score 10 · Accepted Answer · answered Dec 01 '15 at 00:15

Continuing on with my comment, you can use a dict comprehension:

TD = {key:[] for key in hw}

would create an empty list for each item in hw. For example:

>>> hw = ['h1','h2']
>>> TD = {key:[] for key in hw}
>>> TD
{'h1': [], 'h2': []}
>>> TD['h1'].append(1)
>>> TD
{'h1': [1], 'h2': []}

To answer your original question:

The reason that changing one list would have changed every list is because dict.fromkey(seq,[value]) is a function that creates a new dictionary where every item in seq will map to the optional argument value. You passed in an empty list as value which means that you set each key to point to a reference to the same list. Since all the keys point to the same list, then obviously changing the list by accessing ANY key (try accessing h2 and you will see this is true) will reflect that change across all keys.

score 1 · Answer 2 · answered Dec 01 '15 at 00:14

1

You can do this:

TD = dict((i, [],) for i in hw)

answered Dec 01 '15 at 00:14

sisanared

4,175
2
27
42

score 1 · Answer 3 · answered Dec 01 '15 at 00:18

The magic is happening in TD = TD.fromkeys(hw, []).

It's making a single list, and assigning that one list as the value to both of your keys. Thus, when you change either one key, you're changing the single internal list, and thus changing what both keys are pointing at.

To illustrate the point, check out:

row_of_words = ['word 1', 'word 2']
hw = ['h1', 'h2']
TD = {}
A = []

TD = TD.fromkeys(hw, A)

for index in range(len(row_of_words)):
    TD['h1'].append(1)
    print index, TD

print TD  # {'h2': [1, 1], 'h1': [1, 1]}
A.append(2)
print TD  # {'h2': [1, 1, 2], 'h1': [1, 1, 2]}

score 0 · Answer 4 · answered Jun 23 '21 at 07:50

The answers apply to both, Python 3 and 2. Just wanted to provide two comparable examples with the methods provided here.

Problem

d = dict.fromkeys([1,2,3],[]) # list to empty dict, method 1
d[1].append("foo") # assumingly append item to key 1 only
d
# {1: ['foo'], 2: ['foo'], 3: ['foo']} # appends to all keys instead

Solution

d = {key:[] for key in [1,2,3]} # list to empty dict, method 2
d[1].append("foo") # really only append item to key 1
d
# {1: ['foo'], 2: [], 3: []} # works

Appending to one python dictionary key appends to all for some reason

4 Answers4

Problem

Solution