Spark Neo4j Create/Merge Relationship based on relationship key

Neo4j Graph Platform Import / Export
1

I would like to create/overwrite relationships between two nodes based on relationship key by using Neo4J Spark. I understand that Neo4J Spark provide source node key & target node key but do not have relationship key which i can't create/overwrite relationships between two nodes based on relationship key.

Below is the sample data model:

image
image846×312 6.7 KB

Scenario
Source database
row 1: Customer A bought a Product: Pen by quantity: 10 at 25 Sept 2024
row 2: Customer A bought a Product: Pen by quantity: 15 at 26 Sept 2024
Both row import into Neo4J Graph with relationship: HAS_BOUGHT, properties: quantity, transaction_date

Source database updated row 1 quantity from 10 to 12
row 1: Customer A bought a Product: Pen by quantity: 12 at 25 Sept 2024
row 2: Customer A bought a Product: Pen by quantity: 15 at 26 Sept 2024

I can't update row1 using Spark Neo4j into Neo4J Graph since there are no relationship key to specify for relationship update. i understand we can add transaction node as composite node between Customer & Product but i don't want make the data model complex. I need relationship as transaction records.

Thanks

2

I suggest you don't use the relationship to store the transaction information. I suggest you use a "Transaction" node to store the transaction properties. Then the transaction will have a relationship ("HAS_CUSTOMER") to the customer and one or more relationships to products via a "HAS_PURCHASED" relationships.

3

Gary's answer is the preferred course of action for the time being.
I just filed a new issue to our team so that we introduce support for relationship keys as well as we currently do for node keys. No timeline for when/if it becomes available however.

4

Thanks to Gary & Florent. Understand that with composite node would be the preferred action but i have to follow this data model. For the time being, i would use custom cypher query to ingest the relationship.