Transforming model instances into other formats such as JSON or XML is a core capability needed when exposing APIs or exchanging data between systems. In Django, the built-in django.core.serializers module provides a straightforward way to convert querysets into serialized representations with out requiring extra dependencies.

Basic Usage

To produce a serialized output, import the module and call serialize() with the desired format name and a queryset. Valid formats include json and xml.

from django.core import serializers

records = serializers.serialize("json", Author.objects.all())
print(records)

The returned string contains a list of objects, each holding the model identifier, primary key, and a fields dictionary of attribute values.

[
  {
    "model": "index.author",
    "pk": 2,
    "fields": {
      "name": "pfeiliu",
      "age": 19,
      "email": "789@163.com",
      "is_active": false
    }
  }
]

Limiting the Exported Fields

When you only need a subset of a model’s fields, provide the fields keyword argument as a tuple of attribute names:

selection = serializers.serialize(
    "xml", 
    Author.objects.only("name", "age"), 
    fields=("name", "age")
)
print(selection)

For JSON output, the same approach applies:

partial_data = serializers.serialize(
    "json", 
    Author.objects.all(), 
    fields=("name", "age")
)
print(partial_data)

The resulting structure still includes model and pk, but fields will contain only the keys specified.

[
  {
    "model": "index.author",
    "pk": 2,
    "fields": {
      "name": "pfeiliu",
      "age": 19
    }
  }
]

Deserialization and Reconstructing Objects

Serialization is only one direction. To recreate model instances from a serialized stream, use deserialize:

for deserialized_object in serializers.deserialize("json", json_data):
    deserialized_object.save()

Each iteration yields a DeserializedObject that wraps the model instance, allowing bulk operations or individual saving.