The current behaviour of the ingester is to read the index file and create two files:
Objects, which contain all the vertices.
Relationships, which contain all the relationships between Objects.
Then, we have to manually import these files into ADB.
The correct behaviour is that the ingester ingests all data and takes care of ingestions. There are cases which we should take care of:
There are unique constraints. Therefore, we cannot re-ingest data. Therefore, for development, there should be a flag that lets us do it. It can be done by passing overwrite=True to the database. It should not be the default behaviour.
We must not ingest relationships if the vertices ingestion fails.
The current behaviour of the ingester is to read the index file and create two files:
Objects, which contain all the vertices.Relationships, which contain all the relationships between Objects.Then, we have to manually import these files into ADB.
The correct behaviour is that the ingester ingests all data and takes care of ingestions. There are cases which we should take care of:
overwrite=Trueto the database. It should not be the default behaviour.