Ndtyjky

Using Django 2.1/ python 3.6 I'm looking to update a field on an existing model.

The field isof the following type:

from django.contrib.postgres.fields import (

    ArrayField,

    JSONField

)

waypoints = ArrayField(

    JSONField(

        default=list,

        null=True,

        blank=True

    ),

    size=None,

    null=True,

    blank=True

)

And the field will store data in the format:

[

    {'lat': 63.123334, 'lon': 13.433918, 'date': '2018-08-23 11:00:00', 'direction': 123, 'rpm': 0},

    {'lat': 42.315119, 'lon': -3.213883, 'date': '2018-08-12 09:15:00', 'direction': 95.45, 'rpm': 3998},

    {'lat': 51.763023, 'lon': 7.376109, 'date': '2018-08-19 03:30:00', 'direction': 45.76, 'rpm': 7823}

]

I am acutely aware of the fact that this is not good practice, and that the better method would be to use a foreign key with another model, but my hands are tied.

My code is:

scheduler_response = {

    "result": {

        "time": 23,

        "total_consumption": 98117,

        "waypoints": [

            {

                "lat": 42.315119,

                "lon": -3.213883,

                "date": "2018-08-12 09:15:00",

                "direction": 95.45,

                "rpm": 3998,

                "cumulative_consumption": 0

            },

            {

                "lat": 51.763023,

                "lon": 7.376109,

                "date": "2018-08-19 03:30:00",

                "direction": 45.76,

                "rpm": 7823,

                "cumulative_consumption": 44298

            },

            {

                "lat": 63.123334,

                "lon": 13.433918,

                "date": "2018-08-23 11:00:00",

                "direction": 123,

                "rpm": 0,

                "cumulative_consumption": 98117

            }

        ]

    }

}

scheduler_waypoints = scheduler_response.get('result').get('waypoints')

job_updates = {}

job_updates['waypoints'] = scheduler_waypoints

print(type(job_updates['waypoints']))

job_updates['total_consumption'] = scheduler_response.get('total_consumption')

serialized_job = JobSerializer(active_job, data=job_updates, partial=True)



if serialized_job.is_valid():

    print('is_valid')

    print(serialized_job.validated_data)

    serialized_job.save()

    print('FINISHED')

My terminal output is as follows:

<class 'list'>

is_valid

OrderedDict([('total_consumption', None), ('waypoints', [{'lat': 42.315119, 'lon': -3.213883, 'date': '2018-08-12 09:15:00', 'direction': 95.45, 'rpm': 3998}, {'lat': 51.763023, 'lon': 7.376109, 'date': '2018-08-19 03:30:00', 'direction': 45.76, 'rpm': 7823}, {'lat': 63.123334, 'lon': 13.433918, 'date': '2018-08-23 11:00:00', 'direction': 123, 'rpm': 0}])])

Note that although the serializer is validated, the serializer cannot be saved ('FINISHED' is not printed).

I am getting the error:

Traceback (most recent call last):

File "/usr/local/lib/python3.6/site-packages/django/db/backends/utils.py", line 85, in _execute

    return self.cursor.execute(sql, params)

psycopg2.ProgrammingError: column "waypoints" is of type jsonb but expression is of type text

LINE 1: ...d" = 1, "total_consumption" = NULL, "waypoints" = ARRAY['{"l...

                                                            ^

HINT:  You will need to rewrite or cast the expression.

Why is this? What am I doing wrong (aside from not using a foreign key and another model)?

The answer for anyone else looking to commit this kind of attrocity is to use a JSONBField in your model, not a JSONField inside a list field, here are the changes i made to my model:

from django.contrib.postgres.fields.jsonb import JSONField as JSONBField

class Job(models.Model):

    waypoints = JSONBField(

        default=list,

        null=True,

        blank=True

    )

The code calling the serializer then works fine.