Python - comparing JSON data structures.

README.md

Python - comparing JSON data structures

A function compare_json_data(source_data_a,source_data_b), accepting structures populated with data loaded from json.load() and comparing for equality.

Example

$ ./compare.py 
Compare JSON result is: True

JSON files a.json and b.json are loaded via load_json() function and structures passed into compare_json_data() for comparison.

a.json

{
	"apple": "value",
	"orange": 12,
	"banana": [1,2,3,"One","Two","Three","Fourth"],
	"grape": false,
	"strawberry": null,
	"carrot": [
		{
			"fourth": "value-two",
			"first": {
				"apple": "value1",
				"orange": "value2"
			},
			"second": "value",
			"third": "another"
		},
		{},
		["One","Two","Three"]
	],
	"pear": {
		"first": "value",
		"second": "another",
		"third": [1,2,3,"One","Two","Three"],
		"fourth": [
			{
				"first": "value",
				"second": "another"
			},
			{}
		]
	}
}

b.json

{
	"apple": "value",
	"orange": 12,
	"banana": [1,2,3,"One","Two","Three","Fourth"],
	"pear": {
		"first": "value",
		"second": "another",
		"third": [1,2,3,"One","Two","Three"],
		"fourth": [
			{
				"second": "another",
				"first": "value"
			},
			{}
		]
	},
	"grape": false,
	"strawberry": null,
	"carrot": [
		{
			"second": "value",
			"third": "another",
			"fourth": "value-two",
			"first": {
				"apple": "value1",
				"orange": "value2"
			}
		},
		{},
		["One","Two","Three"]
	]
}

compare.py

#!/usr/bin/env python3

import json


def load_json(file_path):
    # open JSON file and parse contents
    fh = open(file_path, "r")
    data = json.load(fh)
    fh.close()

    return data


def compare_json_data(source_data_a, source_data_b):
    def compare(data_a, data_b):
        # type: list
        if type(data_a) is list:
            # is [data_b] a list and of same length as [data_a]?
            if (type(data_b) is not list) or (len(data_a) != len(data_b)):
                return False

            # iterate over list items
            for list_index, list_item in enumerate(data_a):
                # compare [data_a] list item against [data_b] at index
                if not compare(list_item, data_b[list_index]):
                    return False

            # list identical
            return True

        # type: dictionary
        if type(data_a) is dict:
            # is [data_b] a dictionary?
            if type(data_b) is not dict:
                return False

            # iterate over dictionary keys
            for dict_key, dict_value in data_a.items():
                # key exists in [data_b] dictionary, and same value?
                if (dict_key not in data_b) or (
                    not compare(dict_value, data_b[dict_key])
                ):
                    return False

            # dictionary identical
            return True

        # simple value - compare both value and type for equality
        return (data_a == data_b) and (type(data_a) is type(data_b))

    # compare a to b, then b to a
    return compare(source_data_a, source_data_b) and compare(
        source_data_b, source_data_a
    )


def main():
    # import testing JSON files to Python structures
    a_json = load_json("a.json")
    b_json = load_json("b.json")

    # compare first struct against second
    print(f"Compare JSON result is: {compare_json_data(a_json, b_json)}")


if __name__ == "__main__":
    main()

I'm sorry, but it's not same that if '==' betwen two json? I think that would be better return what key or value is different.

you might want to sort the lists before comparing them, and length of dictionaries could be compared as well

Yeah, Sort the 2 json objects by keys and then just apply ==
Wouldn't that be a simple solution ?

@avoidik if I sort the lists, they are no longer considered equal... e.g.

[1,2,3,"One","Two","Three"]
[1,2,3,"One","Three","Two"]

those are not equal - what I want... but if sorted first before index/value compare...

[1,2,3,"One","Two","Three"]
[1,2,3,"One","Two","Three"]

now they are equal... but not from the source JSON struct.

Actually my response will return with timestamp. In such cases, it is failing as the timestamp differes. could you please help me to handle those scenarios?

Sorry @ManikandanRajendran you'll have to explain this with a code sample.

If the list contains different no of elements, then the code is throwing error as :

IndexError: list index out of range.

Hey @prixgoody - that's weird - this check should mean that's not possible:

(len(data_a) != len(data_b))

do you have an example of data structures that throw this error?

Hey @prixgoody - that's weird - this check should mean that's not possible:

(len(data_a) != len(data_b))

do you have an example of data structures that throw this error?

Hi, I have used your code but instead of returning true false, i am fetching more details like for which value it is diff. Still in progress.
Sorry for the miscommunication caused.

Hey @prixgoody - that's weird - this check should mean that's not possible:

(len(data_a) != len(data_b))

do you have an example of data structures that throw this error?

Hi, I have used your code but instead of returning true false, i am fetching more details like for which value it is diff. Still in progress.
Sorry for the miscommunication caused.

Any luck fetching which value is diff? I'm working on it too.