Create new nested dictionary using nested key, values

Question

I have a dictionary of key-value pairs, where the value is itself a dictionary. I would like to change the names of key values in that nested dictionary based on a predefined conversion.

I am using two lists to match up the values of the nested keys that I am trying to conver (Old_item1 should become New_item1):

comparison_list = ['Old_item1', 'Old_item2']
new_prods = ['New_item1', 'New_item2']

old_dict = {
    'Company1':
        {
            'Old_item1':
                {
                    'key1': val,
                    'key2': val
                },
            'Old_item2':
                {
                    'key1': val,
                    'key2': val
                }
        }
}

I tried this:

new_dict = {}
for i in comparison_list:
    for j in new_prods:
        new_dict['Company1'][j] = test['Company1'][i]

I get a KeyError: KeyError: 'Company1'

The desired output for each item I add to each list is:

new_dict = {
    'Company1':
        {
            'New_item1':
                {
                    'key1': val  # old item key, val
                    'key2': val  # old item key, val
                }
        }
}

You will probably want as an intermediate step to set up a dictionary that maps the old keys to the new keys (use zip on the two lists to iterate when constructing this). — alani
– alani, Commented Dec 2, 2021 at 18:48
@chasedcribbet Nonsense! You're plenty smart - I hope the edit helps teach the widely used language for this sort of thing, but I also want to be sure you don't feel the need to apologize for asking a question or being new to this particular area of knowledge. That's why we're all here! — Nathaniel Ford
– Nathaniel Ford, Commented Dec 2, 2021 at 19:13
@OlavAga made a good point below. The question now looks like I wanted to change the names, but what I wanted is a new, separate dict. — chasedcribbet
– chasedcribbet, Commented Dec 2, 2021 at 20:23

alani · Accepted Answer · 2021-12-02 18:58:56Z

3

You can make a dictionary of mappings from the old to the new items, and then use it to create a new sub-dictionary for each company. This then needs to be wrapped inside an outer loop over companies (although here there is only one).

For example:

comparison_list = ['Old_item1', 'Old_item2']
new_prods = ['New_item1', 'New_item2']
    
old_dict = {'Company1': 
            {'Old_item1': 
             {'key1': 2, 
              'key2': 3},
             'Old_item2': 
             {'key1': 4,
              'key2': 5}}} 

key_mappings = dict(zip(comparison_list, new_prods))

new_dict = {k: {key_mappings[k1]: v1 for k1, v1 in v.items()}
            for k, v in old_dict.items()}

print(new_dict)

gives:

{'Company1': {'New_item1': {'key1': 2, 'key2': 3}, 'New_item2': {'key1': 4, 'key2': 5}}}

Here is the mappings dictionary key_mappings which we used:

{'Old_item1': 'New_item1', 'Old_item2': 'New_item2'}

edited Dec 2, 2021 at 18:58

answered Dec 2, 2021 at 18:52

alani

13.2k3 gold badges18 silver badges34 bronze badges

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

4 Comments

Nathaniel Ford Over a year ago

Be aware with this solution that it will not handle any keys in the old_dict that are not also in the key_mapping and give you a KeyError.

alani Over a year ago

@NathanielFord Yes, that is correct. I was assuming that all the keys will be in those lists.

alani Over a year ago

@NathanielFord If the keys are not present, then I guess there are different options depending on the requirements. It might be that the original key should be used - in which case a get as you suggest (here: key_mappings.get(k1, k1) in place of key_mappings[k1]) will do the job nicely. Or it might be that it should be omitted entirely, in which case an if k1 in key_mappings inside the inner dict comprehension could be used.

Nathaniel Ford Over a year ago

I agree! It's not specified in the OP, and there are different dynamics depending on which way you want to go. - which you've described quite well. :)

Nathaniel Ford · Accepted Answer · 2021-12-02 19:03:00Z

2

The easiest way to go about this is to re-create your dictionary with a comprehension:

new_dict = {company: update_item(v) for company, v in old_dict.items()}

Then parse out update_item to it's own function. You could do this inline but it makes it difficult to understand.

conversion_lookup = {
    'Old_item1': 'NewItem1',
    'Old_item2': 'NewItem2',
}

def update_item(item: dict) -> dict:
    return { conversion_lookup.get(k, k): v for k, v in item.items() }

new_dict = {company: update_item(v) for company, v in old_dict.items()}

For the conversion here I'm using a dictionary describing the conversion. If you need to construct this automatedly:

comparison_list = ['Old_item1', 'Old_item2']
new_prods = ['New_item1', 'New_item2']

conversion_lookup = { v: new_prods[idx] for idx, v in enumerate(comparison_list) }

The reason I like a dictionary is that you can use some_dict.get(a_value, a_default_value). Then, if your value isn't in your conversion dictionary you can fall back to the original value. That's what I'm doing here:

conversion_lookup.get(k, k)

The second k is the original item's value, which is a good thing to use if your conversion list doesn't include what you want.

edited Dec 2, 2021 at 19:03

answered Dec 2, 2021 at 18:54

Nathaniel Ford

21.3k20 gold badges98 silver badges112 bronze badges

2 Comments

alani Over a year ago

With your for company, v in old_dict, do you mean for company, v in old_dict.items()? If you iterate over a dictionary, you will just get the keys.

Nathaniel Ford Over a year ago

Yup! I do. :) Thank you!

Olav Aga · Accepted Answer · 2021-12-02 19:27:51Z

0

One approach is to do it step by step. First create the dicts with the new names and old keys, then remove the old keys, and finally add in the new names.

# Keep the values from the old keys
values = [old_dict['Company1'][old_name] for old_name in comparison_list]

# Remove old names, and add in the 
for new_key, value, old in zip(new_prods, values, comparison_list):
    old_dict['Company1'].pop(old)
    old_dict['Company1'][new_key] = value

edited Dec 2, 2021 at 19:27

answered Dec 2, 2021 at 18:55

Olav Aga

14310 bronze badges

7 Comments

Nathaniel Ford Over a year ago

I'm not sure what's being accomplished with the pop? They should simply be overwritten when you do the update, right?

Nathaniel Ford Over a year ago

I think you might be off by one with regards to nesting level. Running your code puts new_item1, etc. on the same level as Company1 - which is not what you want. I think this is an awkward approach for that reason.

alani Over a year ago

To correct @NathanielFord's last point (subject to the existing limitation of the single hard-coded top-level key) just change old_dict.update to old_dict['Company1'].update

Nathaniel Ford Over a year ago

Generally in Python I find it as or more efficient to create a new dict and just replace the previous one - the way the underlying dynamics work it's basically the same thing and way easier to reason about. Also, alani is correct about the fix for the update issue, though you'd have to do it for each company and you'd have to put the whole thing in a loop or something similar to make that work.

chasedcribbet Over a year ago

My original question pointed to a new_dict. I think it may have been lost in the edits.

|

Collectives™ on Stack Overflow

Create new nested dictionary using nested key, values

3 Answers 3

4 Comments

2 Comments

7 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

4 Comments

2 Comments

7 Comments

Your Answer

Sign up or log in

Post as a guest

Related