CODEWITHSUNDEEP
pythonjson
Raoul
Raoul

Reputation: 238

Rename JSON key names

I receive a JSON object like this: 

{
    "Question Communicating": "Natural language",
    "interpretation_type": "recognition",
    "output1": "test",
    "Question Learning": "Reinforcement",
    "output2": "test2",
    "output3": "something"
}

My question is, is it possible to rename the key name: 'outputX' to 'output'. I don't know how many times 'outputX' will be in the JSON, but I need all the outputs renamed to 'output'. So it will end up like this:

{
    "Question Communicating": "Natural language",
    "interpretation_type": "recognition",
    "output": "test",
    "Question Learning": "Reinforcement",
    "output": "test2",
    "output": "something"
}

Upvotes: 2

Views: 596

Answers (2)

chepner
chepner

Reputation: 530930

One possibility is to use a data structure that allows duplicate keys, such as webob.multidict.Multidict.

import webob.multidict
import json


class MultiDictEncoder(json.JSONEncoder):
    def default(self, o):
        if isinstance(o, webob.multidict.MultiDict):
            return o
        else:
            return super().default(o)

    def encode(self, o):
        if isinstance(o, webob.multidict.MultiDict):
            # Just a proof of concept. No attempt is made
            # to properly encode keys for values.
            return ('{'
                    + ', '.join(f'"{k}": "{v}"' for k, v in o.items())
                    + '}')
        else:
            return super().encode(o)


with open("tmp1.json") as f:
    input_data = json.load(f)

output_data = webob.multidict.MultiDict()
for k, v in input_data.items():
    if k.startswith("output"):
        k = 'output'
    output_data.add(k, v)

with open("tmp2.json", 'w') as f:
    print(json.dumps(output_data, cls=MultiDictEncoder), file=f)

For some reason in testing this, using json.dump produced an error involving circular references. I don't know if this is a problem with how I defined MultiDictEncoder.default, but the resulting tmp2.json does have duplicate output keys.

Upvotes: 0

RoadRunner
RoadRunner

Reputation: 26315

Trying to use duplicate keys in a JSON object is not recommended. You can see the problems that arise when you serialize and deserialize duplicate keys, or try to force them into a dictionary. The duplicate keys are not retained.

>>> from json import dumps, loads
>>> json = '{"a": "x", "a": "y"}'
>>> loads(json)
{'a': 'y'}
>>> json = {'a': 'x', 'a': 'y'}
>>> dumps(json)
'{"a": "y"}'
>>> json = {'a': 'x', 'a': 'y'}
>>> json
{'a': 'y'}

Instead you could try grouping all keys that start with "output" into a list ["test", "test2", "something"].

from json import dumps

d = {
    "Question Communicating": "Natural language",
    "interpretation_type": "recognition",
    "output1": "test",
    "Question Learning": "Reinforcement",
    "output2": "test2",
    "output3": "something"
}

result = {}
for k, v in d.items():
    if k.startswith("output"):
        result.setdefault("output", []).append(v)
    else:
        result[k] = v

print(dumps(result, indent=4))

Output JSON:

{
    "Question Communicating": "Natural language",
    "interpretation_type": "recognition",
    "output": [
        "test",
        "test2",
        "something"
    ],
    "Question Learning": "Reinforcement"
}

Upvotes: 2

Related Questions