#4347: Changelog & cleanup

Author: jeremystretch

docs/release-notes/version-3.4.md

@@ -17,6 +17,10 @@
 
 NetBox's global search functionality has been completely overhauled and replaced by a new cache-based lookup.
 
+### JSON/YAML Bulk Imports ([#4347](https://github.com/netbox-community/netbox/issues/4347))
+
+NetBox's bulk import feature, which was previously limited to CSV-formatted data for most objects, has been extended to support the import of objects from JSON and/or YAML data as well.
+
 #### CSV-Based Bulk Updates ([#7961](https://github.com/netbox-community/netbox/issues/7961))
 
 NetBox's CSV-based bulk import functionality has been extended to support also modifying existing objects. When an `id` column is present in the import form, it will be used to infer the object to be modified, rather than a new object being created. All fields (columns) are optional when modifying existing objects.

netbox/netbox/views/generic/bulk_views.py

@@ -375,19 +375,14 @@ def create_and_update_objects(self, form, request):
                     form.add_error('data', f"Row {i}: Object with ID {object_id} does not exist")
                     raise ValidationError('')
 
+            # Instantiate the model form for the object
+            model_form_kwargs = {
+                'data': record,
+                'instance': instance,
+            }
             if form.cleaned_data['format'] == ImportFormatChoices.CSV:
-                model_form = self.model_form(record, instance=instance, headers=form._csv_headers)
-            else:
-                model_form = self.model_form(record, instance=instance)
-                # Assign default values for any fields which were not specified.
-                # We have to do this manually because passing 'initial=' to the form
-                # on initialization merely sets default values for the widgets.
-                # Since widgets are not used for YAML/JSON import, we first bind the
-                # imported data normally, then update the form's data with the applicable
-                # field defaults as needed prior to form validation.
-                for field_name, field in model_form.fields.items():
-                    if field_name not in record and hasattr(field, 'initial'):
-                        model_form.data[field_name] = field.initial
+                model_form_kwargs['headers'] = form._csv_headers
+            model_form = self.model_form(**model_form_kwargs)
 
             # When updating, omit all form fields other than those specified in the record. (No
             # fields are required when modifying an existing object.)