How can I merge two tables without label in spark?

I'm working on ETL with neo4j in Oracle on Spark right now.
The language is scala.

Just curious, I have two different tables.
However, when merging the two tables, I want to add only a relationship without creating a label or node on one side.

For example, we have tables table1 , table2 .
Table1 contains customer information, and table2 contains information about people I like.
To simplify this with a cypher statement:

match (a: tabel1. Customer_id),(b: table1.Customer_id) where a.customer_id =table2.Customer_id and table1.Customer_id =table2.follow_id
merge (a)-[r:INTERESTED_IN]->(b)

This will create only additional nodes that do not have a non-existent relationship.
How can I code this in spark in scala?