Django import export: update without add - django

Question

I'm using 'django import export' (DIE) for importing and updating some data.
Import process starts from checking exists objects in DB, searching by values in ID-field, and if row with ID from import file not found - new entre will be created. How can i made "update only" scenario, where if 'id key' not found in DB, row will be skipped (not added new)?
my model.py
class Size(models.Model):
id = models.AutoField(unique=True, primary_key=True, null=False, blank=False)
height = models.SmallIntegerField()
width = models.SmallIntegerField()
class Product(models.Model):
id = models.AutoField(unique=True, primary_key=True, null=False, blank=False)
vendor_code = models.CharField(unique=True, max_length=50, null=False, blank=False)
price = models.DecimalField(null=False, blank=False)
size = models.ForeignKey(Size, verbose_name=u'Size')
in resource.py
class ProductSyncResource(resources.ModelResource):
class Meta:
model = ProductVariant
import_id_fields = ('vendor_code',)
fields = ('vendor_code', 'price',)
export_order = ('vendor_code', 'price', 'status', )
skip_unchanged = True
report_skipped = True
dry_run = True
import table (xls)
If vendor_code 'Tк-12856' (Cell A3) will be not found, then DIE will try to add this row, and:
We will get error from DB (foreignKey check for column 'size')
I don't need to add this row to DB in my 'update scenario'

Answer 1

Finally i got it by overriding skip_row. Fields now can be 'null=False' and will be imported only rows with known import_id_field values.
class VariantSyncResource(resources.ModelResource):
class Meta:
model = ProductVariant
import_id_field = 'vendor_code'
import_id_fields = ('vendor_code',)
fields = ('vendor_code', 'price', 'status', )
export_order = ('vendor_code', 'price', 'status', )
skip_unchanged = True
report_skipped = False
dry_run = True
def skip_row(self, instance, original):
original_id_value = getattr(original, self._meta.import_id_field)
instance_id_value = getattr(instance, self._meta.import_id_field)
if original_id_value != instance_id_value:
return True
if not self._meta.skip_unchanged:
return False
for field in self.get_fields():
try:
if list(field.get_value(instance).all()) != list(field.get_value(original).all()):
return False
except AttributeError:
if field.get_value(instance) != field.get_value(original):
return False
return True

Answer 2

If you want it to not create new objects I think you'd need to override the import_row() within ProductSyncResource.
Then you could just do if new: return;
def import_row(self, row, instance_loader, using_transactions=True, dry_run=False, **kwargs):
row_result = self.get_row_result_class()()
try:
self.before_import_row(row, **kwargs)
instance, new = self.get_or_init_instance(instance_loader, row)
self.after_import_instance(instance, new, **kwargs)
if new:
return row_result
else:
row_result.import_type = RowResult.IMPORT_TYPE_UPDATE
row_result.new_record = new
original = deepcopy(instance)
...
Following your comment about the preview error, just remember that this isn't an official feature of the app so you'll have to just follow stack traces & work around issues that popup. My first thought above looked like the quickest & easiest way to do this, but you could try to make the else for the save save conditional against new objects;
elif not new:
with transaction.atomic():
self.save_instance(instance, using_transactions, dry_run)
self.save_m2m(instance, row, using_transactions, dry_run)
diff.compare_with(self, instance, dry_run)