Extract which required key in JSON is missing from ValidationError

Question

I try to validate JSON in python using jsonschema library. What I want to achieve is a dictionary with keys conststing of json keys causing trouble and messages from validator as values.

There is no problem with type validation as it returns path to json key which does not pass validation. My problem is when key is missing path variable in ValidationError is empty and I can't find any other place where it's stated which json key is causing trouble. The only place that states which json key is the problem is message, but spliting it to extract that information seems more like workaround than solution.

Code:

import jsonschema

schema = {
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "type": "object",
    "properties": {
        "name": {
            "type": "string"
        },
        "email": {
            "type": "string"
        }
    },
    "required": ["name", "email"]
}

jsonToValidate = {
  "name": 1234
}

validator = jsonschema.Draft202012Validator(schema)
errorsGenerator = validator.iter_errors(jsonToValidate)
errors = dict()
for err in errorsGenerator:
    # errors.update({err.path[0], err.message})
    print("path: ", err.path, " \n message: ", err.message)

Current output:

path:  deque(['name'])  
 message:  1234 is not of type 'string'
path:  deque([])  
 message:  'email' is a required property

Expected output: Dictionary with key as json key that is troublesome and message as value. For this example it would look like this:

{
    "name": "1234 is not of type 'string'",
    "email":  'email' is a required property
}

So my question is: Is it possible to extract information which missing json key (that is required by schema) caused ValidationError?

simon · Accepted Answer · 2023-07-24T10:01:34.253

There is an open bug report for exactly your problem in the jsonschema repo, with the short answer that there is currently no standard way for getting the missing required key. However, a comment to the report mentions a workaround via custom validators. Maybe this would be a way to go for you as well.

Update: Alternatively, you can maybe use a mix of message parsing and using available attributes: Parse the error message for required keys, as suggested in user2704177's answer; otherwise, use the ValidationError's path where possible. So, something like:

errors = dict()
for err in errorsGenerator:
    if err.validator == "required":
        # Parse first word of error message as key, strip surrounding quotes
        errors[err.message.split(" ")[0].strip("'")] = err.message
    else:
        # Use last path element as key (empty string for empty path)
        errors[err.path[-1] if err.path else ""] = err.message
print(errors)
# >>> {'name': "1234 is not of type 'string'", 'email': "'email' is a required property"}

Thx for pointing to the bug report. I'll maybe try the workaround you mentioned, but it's not as crucial for now. — MrHutnik, Jul 24 '23 at 21:28

score 0 · Answer 2 · answered Jul 22 '23 at 18:29

0

Based on your example it looks like it's the first word of the error message, so you could extract the first word:

err.message.split(" ")[0]

And use that to generate your dictionary.

answered Jul 22 '23 at 18:29

user2704177

109
2
6

Well yes, but then I'll need to add code to verify if the error refers to missing field or mismatched type and treat them differently. That complicates the code a bit. It's a solution for sure, but I was wondering if there is something more elegant/universal – MrHutnik Jul 22 '23 at 18:34

Extract which required key in JSON is missing from ValidationError

2 Answers2