Use string variable as key in nested json in Python

Question

I have a JSON object, data, I need to modify.

Right now, I am modifying the object as follows:

data['Foo']['Bar'] = 'ExampleString'

Is it possible to use a string variable to do the indexing?

s = 'Foo/Bar'
data[s.split('/')] = 'ExampleString'

The above code does not work. How can I achieve the behavior I am after?

NB : I am looking for a solution which supports arbitrary number of key "levels", for instance the string variable may be Foo/Bar/Baz, or Foo/Bar/Baz/Foo/Bar, which would correspond to data['Foo']['Bar']['Baz'] and data['Foo']['Bar']['Baz']['Foo']['Bar'].

@shaikmoeed Yes, but what if the number of elements is variable? E.g. Foo/Bar/Baz — Paolo
– Paolo, Commented Feb 3, 2020 at 11:28
does your string get a bit complex? I mean like 'Foo/Bar/inner/first/almost/here'? — Ramesh Reddy
– Ramesh Reddy, Commented Feb 3, 2020 at 11:29
@Ramesh Yes, the solution should supported arbitrary number of fields — Paolo
– Paolo, Commented Feb 3, 2020 at 11:29
Then, you should go with a loop until you reach the last element. Is that what you are looking for? — shaik moeed
– shaik moeed, Commented Feb 3, 2020 at 11:32

kabanus · Accepted Answer · 2020-02-03 12:09:34Z

1

Without changing completely the data class you want to use, this might be easisest:

def jsonSetPath(jobj, path, item): 
    prev = None 
    y = jobj
    for x in path.split('/'):
        prev = y
        y = y[x]
    prev[x] = item

A wrapper Python to descend iteratively into the object. Then you can use

jsonSetPath(data, 'foo/obj', 3)

normally. You can add this functionality to your dictionary by inheriting dict if you prefer:

class JsonDict(dict):
    def __getitem__(self, path):
        # We only accept strings in this dictionary
        y = self
        for x in path.split('/'):
            y = dict.get(y, x)
        return y

    def __setitem__(self, path, item):
        # We only accept strings in this dictionary
        y = self
        prev = None
        for x in path.split('/'):
            prev = y
            y = dict.get(y, x)
        prev[x] = item

note using UserDict from collections may be advised, but seems to much of a hassle without converting all the inner dictionaries to user dictionaries. Now you wrap your data (data = JsonDict(data)) and use it as you wanted. If you want to use non-strings as your keys, you need to handle that (though I am not sure that makes sense in this specific dictionary implementation).

Note only the "outer" dictionary is your custom dictionary. If the use case is more advanced you would need to convert all the inner ones as well, and then you might as well use UserDictionary.

edited Feb 3, 2020 at 12:09

answered Feb 3, 2020 at 11:35

kabanus

26.3k7 gold badges48 silver badges79 bronze badges

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

6 Comments

DeepSpace Over a year ago

Subclassing UserDict is an interesting idea. Then this behavior can be baked into __getitem__ so no special method is needed

kabanus Over a year ago

@DeepSpace I am writing that asws :)

DeepSpace Over a year ago

BTW, jsonGetPath will not work. You have to verify that you return the dict, otherwise you will return the value. ie you should only do y = y[x] if y[x] is a dict

kabanus Over a year ago

@DeepSpace I don't care which I return. Why should I? If path to a value is given, it's on the caller, isn't it?

DeepSpace Over a year ago

If you consider data to be {'Foo': {'Bar': {'Baz': None}}}, how would you use jsonGetPath to change None to 1?

|

DeepSpace · Accepted Answer · 2020-02-03 11:33:33Z

1

A very naive solution to get you going in the correct direction.

You need to add error handling, for example what happens if somewhere a long the path a key is missing? You can either bail out or add a new dict on the fly.

def update(path, d, value):
    for nested_key in path.split('/'):
        temp = d[nested_key]
        if isinstance(temp, dict):
            d = d[nested_key]
    d[nested_key] = value

one_level_path = 'Foo/Bar'
one_level_dict = {'Foo': {'Bar': None}}

print(one_level_dict)
update(one_level_path, one_level_dict, 1)
print(one_level_dict)

two_level_path = 'Foo/Bar/Baz'
two_level_dict = {'Foo': {'Bar': {'Baz': None}}}

print(two_level_dict)
update(two_level_path, two_level_dict, 1)
print(two_level_dict)

Outputs

{'Foo': {'Bar': None}}
{'Foo': {'Bar': 1}}
{'Foo': {'Bar': {'Baz': None}}}
{'Foo': {'Bar': {'Baz': 1}}}

answered Feb 3, 2020 at 11:33

DeepSpace

82.1k12 gold badges119 silver badges166 bronze badges

1 Comment

Paolo Over a year ago

Thank you for your answer. I decided to go with jsonSetPath from @kabanus as it is slightly more readable in my opinion.

Ramesh Reddy · Accepted Answer · 2020-02-03 12:31:51Z

1

Using recursion:

x = {'foo': {'in':{'inner':9}}}
path = "foo/in/inner";

def setVal(obj,pathList,val):
    if len(pathList) == 1:
        obj[pathList[0]] = val
    else:
     return setVal(obj[pathList[0]],pathList[1:],val)

print(x)
setVal(x,path.split('/'),10)
print(x)

edited Feb 3, 2020 at 12:31

answered Feb 3, 2020 at 12:18

Ramesh Reddy

10.7k3 gold badges22 silver badges38 bronze badges

Collectives™ on Stack Overflow

Use string variable as key in nested json in Python

3 Answers 3

6 Comments

1 Comment

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

6 Comments

1 Comment

Comments

Your Answer

Sign up or log in

Post as a guest

Related