This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author jacob.underwood
Recipients jacob.underwood
Date 2020-05-16.05:02:29
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1589605349.29.0.839939005018.issue40639@roundup.psfhosted.org>
In-reply-to
Content
I was experimenting with nested dictionaries when I came across strange behavior that I can not figure out at all. The following function, at least for me, has some weird behavior. Originally, it was a bit longer, but somehow it achieved its intended purpose in a strange way, allowing me to cut off the rest of the correct code and to leave only the following with the code still continuing to work:

My original intention was to make a function that would use the provided keys (in the "keys" argument of the following function) in the form of a list to use all but the last key in the list to get to a nested dictionary within the provided dictionary arg. Then, using the last key in the list, set the key to the information provided. Finally, the function should have returned the full dictionary originally provided, but with the one change within one of the nested dictionaries of the provided dictionary.

def input_into_nested_dict(dictionary, keys, information):
    its = information
    cdip = dictionary
    for key in keys[:-1]:
            cdip = cdip[key]
    cdip[keys[-1]] = its
    return dictionary

So, an example dictionary of
m = {'a':{'b':{'c':{'d':30}}}}

An example list of
y = ['a','b','c','test']

Test of
n = input_into_nested_dict(m, y, 'hello')

The strange thing is, n now correctly has
{'a': {'b': {'c': {'d': 30, 'test': 'hello'}}}}
which does not correspond to what this weird cut off function should do, as it should only provide
{'d': 30, 'test': 'hello'}

Furthermore, somehow this changes the global m, the dictionary provided to the function, making it
{'a': {'b': {'c': {'d': 30, 'test': 'hello'}}}}


(The strange variable names are left over from my original code, though removing them as they are now unnecessary changes the behavior again, even though I am pretty sure it should not (?))

Same function without the beginning variables:

def b(dictionary, keys, information):
    for key in keys[:-1]:
            dictionary = dictionary[key]
    dictionary[keys[-1]] = information
    return dictionary

Now, when ran with the same m (reset back to the original dictionary before it was somehow changed), y, and an n of
n = b(m, y, 'hello')
The result is now that n gets the correct dictionary it should get
{'d': 30, 'test': 'hello'}

Though, somehow m changing also continues, with it again getting
{'a': {'b': {'c': {'d': 30, 'test': 'hello'}}}}

If this makes no sense, I'm sorry I am new to Python and am not really good really giving this kind of information

In addition to that, this is probably not even a bug at all and I am probably just missing something big? If true, I'm sorry about this post...

I do not know if this happens to just me, this originally happened in 3.7.7 but I tried upgrading and it continued in the most recent version (3.8.3)

Thank you!
History
Date User Action Args
2020-05-16 05:02:29jacob.underwoodsetrecipients: + jacob.underwood
2020-05-16 05:02:29jacob.underwoodsetmessageid: <1589605349.29.0.839939005018.issue40639@roundup.psfhosted.org>
2020-05-16 05:02:29jacob.underwoodlinkissue40639 messages
2020-05-16 05:02:29jacob.underwoodcreate