[SOLVED] Cassandra issue with delete between filter() and save()

Cassandra issue with delete between filter() and save()

I have an issue where a delete between loading an item and saving it back to cassandra causes columns which were not updated to be deleted.

class TestItem(Model):
    uuid = columns.Integer(primary_key=True)
    data_1 = columns.Text()
    data_2 = columns.Text()

TestItem.create(uuid=1, data_1='data_1', data_2='data_2').save()

item = TestItem.filter(uuid=1).first()
TestItem.objects(uuid=1).delete()
item.data_2 = 'data_2_deleted'

item_b.save()
# stdout:
# TestItem(uuid=1, data_1=u'data_1', data_2='data_2_deleted')

TestItem.filter(uuid=1).first()
# stdout:
# TestItem(uuid=1, data_1=None, data_2=u'data_2_deleted')

I would expect data_1 to still be present.

Is this intended behaviour? Can I change my code somehow that either the item is completely deleted or saved with all values that were queried before.

Solution

The issue was that the cassandra driver only writes back the rows which were actually changed.

In my use case I wanted to write all rows, changed or not, so I used:

TestItem.create(**dict(item_b)).save()