Hello,

I am seeking clarification regarding ticket #5929, which is now marked as a duplicate of the recently closed ticket #373. Our work on ticket #5929 involves the development of a generalized MultiColumnField—including creating the field, generating migrations, and setting up database columns using a specific naming convention—with the IP address use case as an example.

I have a few questions:

1. Why was ticket #5929 marked as a duplicate of ticket #373, given that both tickets share similar objectives but target different applications?
2. Should we continue with our current implementation on ticket #5929, or redirect our focus to the discussions and solutions proposed in ticket #373?
3. Are there any specific directives from the Django team to ensure our contribution for ticket #5929 remains aligned with Django's overall goals?

We are eager to contribute meaningfully to Django and would appreciate any guidance or recommendations.

Below are the code snippets and migration details related to our work on ticket #5929:

Example of a model:

PYTHONCopied
from django.db import models
from django.utils import timezone

class Item(models.Model):
    name = models.CharField(max_length=100)
    last_updated = models.DateTimeField(auto_now=True)
    created_at = models.DateTimeField(auto_now_add=True)
    price = models.DecimalField(max_digits=10, decimal_places=2)
    original_price = models.DecimalField(max_digits=10, decimal_places=2)
    stock = models.IntegerField()
    update_interval = models.DurationField(default=timezone.timedelta(days=1))

class UpsertConflict(models.Model):
    number = models.IntegerField(unique=True)
    rank = models.IntegerField()
    name = models.CharField(max_length=15)

class NetworkModel(models.Model):
    ip_address = models.MultiColumnField(sub_fields={ 
        'octet1': models.PositiveIntegerField(),
        'octet2': models.PositiveIntegerField(),
        'octet3': models.PositiveIntegerField(),
        'octet4': models.PositiveIntegerField()
    }, separator='__')

    def __str__(self):
        return f"{self.octet1}.{self.octet2}.{self.octet3}.{self.octet4}"

The implementation of the MultiColumnField class:

PYTHONCopied
class MultiColumnField(Field):
    def __init__(self, *args, **kwargs):
        self.separator = kwargs.pop("separator", ".")
        self.sub_fields = kwargs.pop("sub_fields", {})
        if not isinstance(self.sub_fields, dict):
            raise ValueError("sub_fields must be a dictionary.")

        super().__init__(*args, **kwargs)
        self.sub_field_names = {}  # Initialisé dans contribute_to_class

    def contribute_to_class(self, cls, name, private_only=False, **kwargs):
        super().contribute_to_class(cls, name, private_only=True, **kwargs)
        # Initialisation des sous-champs avec la convention de nommage
        for field_name, field in self.sub_fields.items():
            prefixed_field_name = f"{self.name}_{field_name}"
            self.sub_field_names[field_name] = prefixed_field_name
            field.set_attributes_from_name(prefixed_field_name)
            field.contribute_to_class(cls, prefixed_field_name)

    def get_prep_value(self, value):
        # Préparer les données pour l'insertion dans la base de données
        if isinstance(value, str):
            # Décomposer la valeur en utilisant le séparateur
            parts = value.split(self.separator)
            value = {
                field_name: part for field_name, part in zip(self.sub_fields, parts)
            }
        return {
            self.sub_field_names[field_name]: field.get_prep_value(value[field_name])
            for field_name, field in self.sub_fields.items()
        }

    def from_db_value(self, value, expression, connection):
        # Reconstruire la valeur à partir des sous-champs
        reconstructed_value = self.separator.join(
            str(value[self.sub_field_names[field_name]])
            for field_name in self.sub_fields
        )
        return reconstructed_value

    def get_db_prep_value(self, value, connection, prepared=False):
        # Préparation des données pour la base de données
        return {
            self.sub_field_names[field_name]: field.get_db_prep_value(
                value[field_name], connection, prepared
            )
            for field_name, field in self.sub_fields.items()
        }

    """ def db_type(self, connection):
        # Indique à Django qu'il n'y a pas de colonne de base de données directement associée
        return None """
    
    def get_col(self, alias, output_field=None):
        # Retourner None ou une représentation adaptée pour le champ complexe
        return ''

Migration generated:

PYTHONCopied
# Generated by Django 5.1.dev20231209152046 on 2023-12-10 21:15

from django.db import migrations, models


class Migration(migrations.Migration):

    dependencies = [
        ('myapp', '0002_upsertconflict'),
    ]

    operations = [
        migrations.CreateModel(
            name='NetworkModel',
            fields=[
                ('id', models.BigAutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
                ('ip_address_octet1', models.PositiveIntegerField()),
                ('ip_address_octet2', models.PositiveIntegerField()),
                ('ip_address_octet3', models.PositiveIntegerField()),
                ('ip_address_octet4', models.PositiveIntegerField()),
            ],
        ),
    ]

Database snapshot:

Screenshot_20231211_130840

Thank you in advance for your feedback and guidance.