CODEWITHSUNDEEP
pythonjsonpython-dataclasses
sji gshan
sji gshan

Reputation: 341

How to parse a JSON object into a python dataclass without third party library?

I want to parse json and save it in dataclasses to emulate DTO. Currently, I ahve to manually pass all the json fields to dataclass. I wanted to know is there a way I can do it by just adding the json parsed dict ie. "dejlog" to dataclass and all the fields are populated automactically.

from dataclasses import dataclass,  asdict


@dataclass
class Dejlog(Dataclass):
    PK: str
    SK: str
    eventtype: str
    result: str
    type: str
    status: str

def lambda_handler(event, context):
    try:
        dejlog = json.loads(event['body'])

        x = Dejlog(dejlog['PK'])

        print(x)

        print(x.PK)

Upvotes: 21

Views: 31988

Answers (4)

John Hillery
John Hillery

Reputation: 11

Here is a simple classmethod that handles missing or extra keys. Building on an earlier answer:

import dataclasses
import json

json_data_str = """
{
   "PK" : "Foo",
   "SK" : "Bar",
   "eventtype" : "blah",
   "result" : "something",
   "type" : "badger",
   "status" : "active",
   "unexpected": "I did not expect this"
}
"""

@dataclasses.dataclass
class Dejlog:
    PK: str
    SK: str
    eventtype: str
    result: str
    type: str
    status: str

    @classmethod
    def from_dict(cls, data):
        return cls(
            *[data.get(fld.name)
              for fld in dataclasses.fields(Dejlog)]
        )

json_obj = json.loads(json_data_str)
dejlogInstance = Dejlog.from_dict(json_obj)

Upvotes: 1

Yechiel babani
Yechiel babani

Reputation: 379

You can load json to your dataclass object using hook_object param like this:

import json
from dataclasses import dataclass


@dataclass
class Dejlog:
    PK: str
    SK: str
    eventtype: str
    result: str
    type: str
    status: str


my_obj = json.loads(
    open('/path/to/file', 'r').read(),
    object_hook=lambda args: Dejlog(**args)
)

Upvotes: 0

Chris Burgoyne
Chris Burgoyne

Reputation: 1604

Unpacking works as long as you don't have any unexpected keys in the json object, otherwise you'll get a TypeError. An alternative is to use a classmethod to create instances of the dataclass. Building on the earlier example:

json_data_str = """
{
   "PK" : "Foo",
   "SK" : "Bar",
   "eventtype" : "blah",
   "result" : "something",
   "type" : "badger",
   "status" : "active",
   "unexpected": "I did not expect this"
}
"""

@dataclass
class Dejlog:
    PK: str
    SK: str
    eventtype: str
    result: str
    type: str
    status: str
    
    @classmethod
    def from_dict(cls, data):
        return cls(
            PK = data.get('PK'),
            SK = data.get('SK'),
            eventtype = data.get('eventtype'),
            result=data.get('result'),
            type=data.get('type'),
            status=data.get('status')
        )

json_obj = json.loads(json_data_str)
dejlogInstance = Dejlog.from_dict(json_obj)

Upvotes: 11

Don
Don

Reputation: 274

As mentioned in other comments you can use the in-built json lib as so:

from dataclasses import dataclass
import json

json_data_str = """
{
   "PK" : "Foo",
   "SK" : "Bar",
   "eventtype" : "blah",
   "result" : "something",
   "type" : "badger",
   "status" : "active"
}
"""

@dataclass
class Dejlog:
    PK: str
    SK: str
    eventtype: str
    result: str
    type: str
    status: str


json_obj = json.loads(json_data_str)
dejlogInstance = Dejlog(**json_obj)

print(dejlogInstance)

Upvotes: 16

Related Questions