Result of comparison depends on the order of datasets being compared

[fedorov]

This tool does not detect absolute difference between the 2 datasets.

Example:

1s.json:

{
    "Key1": "1"
}

2s.json:

{
    "Key1": "1",
    "Key2": "1"
}

$ python comparejson.py 2s.json 1s.json      
Reason: Length Mismatch: Expected Length: 2, Actual Length: 1
Expected:
  {
      "Key1": "1",
      "Key2": "2"
  }
Actual:
  {
      "Key1": "1"
  }
Missing keys: [u'Key2']
$ python comparejson.py 1s.json 2s.json
$

Is this a bug or an (undocumented) feature?

In comparison, jsondiff detects difference no matter what is the order:

$ jsondiff 1s.json 2s.json                                                                                               
{"Key2": "2"}%
$ jsondiff 2s.json 1s.json                                                                                               
{"$delete": ["Key2"]}%

[fedorov]

comparejson.py below:

import json, sys
import jsoncompare
import ast


if len(sys.argv) < 3:
  sys.exit('Error')
json1 = json.loads(open(sys.argv[1],'r').read())
json2 = json.loads(open(sys.argv[2],'r').read())

try:
    ignoredKeys += ast.literal_eval(sys.argv[3])
except IndexError:
    pass

json2 = json.loads(open(sys.argv[2],'r').read())

(check,stack) = jsoncompare.are_same(json1,json2,
                                     ignore_list_order_recursively=False,
                                     ignore_missing_keys=False,
                                     ignore_value_of_keys=ignoredKeys)

if not check:
  print(stack)
  sys.exit(1)