Cannot use error.schema_path to index into error.schema when in a subschema

[martinpengellyphillips]

It seems that error.schema_path always refers to a path in the full schema, whilst error.schema may only be a subschema fragment. This makes it hard (impossible?) to use error.schema_path to index into the relevant schema.

from jsonschema import Draft4Validator as Validator

schema = {
    'type': 'object',
    'properties': {
        'a': {
            'type': 'object',
            'properties': {
                'nested': {
                    'type': 'string',
                    'pattern': '^hello$'
                }
            }
        }
    }
}

errors = Validator(schema).iter_errors({'a': {'nested':'world'}})
for error in errors:
    print error.schema_path, error.schema

# deque(['properties', 'a', 'properties', 'nested', 'pattern']) {'pattern': '^hello$', 'type': 'string'}

Note: For my purposes it would be more useful to access the full schema from the error rather than constrict the schema_path.

[gazpachoking]

Oh right, I was a bit wrong in the other ticket. The schema_path is always for the full schema, except in the instance of errors contained in the context property. The idea was that you always have the full schema, as you just passed it into the validator, but the .schema property of an error was the specific subschema it came from, which you don't already have handy access to.

[martinpengellyphillips]

Ok. In my case multiple schemas are being validated against and the errors bubbling up, so I was hoping my API users could use .schema and .schema_path together. I feel it a bit weird that sometimes schema_path can be used and othertimes it can't, though I understand the logic behind that. In my case I don't need the subschema so might just overwrite the .schema property locally or add a separate full_schema attribute.

[Julian]

I am increasingly in agreement that an absolute_schema attribute is a decent idea.