Help needed with Cypher query

Neo4j Graph Platform Cypher
,
1

The query in Kafka Connect aims to achieve two goals:

  1. Assign multiple labels to a node.
  2. Establish relationships between nodes.

Could help me whether the below query satisfy both the needs?

neo4j.topic.cypher.kafka_car = 
WITH event
CALL {
	WITH event
	MERGE (n:Manufacturer {id: event.id})
	ON CREATE SET n += event
	ON MATCH SET n += event
	WITH n
	SET n:Country:Make:Color
	FOREACH ( _ IN CASE WHEN n.Make = 'BMW' THEN [1] ELSE [] END | SET n:Bmw:Germany)
	FOREACH ( _ IN CASE WHEN n.Make = 'Mercedes' THEN [1] ELSE [] END | SET n:Mercedes:Germany)
	FOREACH ( _ IN CASE WHEN n.Make = 'Land Rover' THEN [1] ELSE [] END | SET n:LandRover:England)
	FOREACH ( _ IN CASE WHEN n.Make = 'Tesla' THEN [1] ELSE [] END | SET n:Tesla:America:Black)
	FOREACH (q IN [(q:Manufacturer {ParentName: n.Name})-[*0]-() | q] | MERGE (n)-[:Parent_To]->(q))
	FOREACH (w IN [(w:Engine {Chasis: n.SerialNumber})-[*0]-() | w] | MERGE (n)-[:Has_Engine]->(w))
}

The query intends to assign the labels "Country," "Make," and "Color" to a node by structuring the data as "Country:Make:Color." Also is this the optimised query?

Additionally, I need to remove any existing relationships for a node before establishing the current relationship, ensuring only the latest relationship is retained. How do I achieve it with the above query?

2

Note, when the make is either of the four car makes, you are setting the labels as specified and removing the three set earlier: Country, Make, and Color.

What is your goal with the last two forEach clauses? The way the pattern is written with a zero path length you will always get the q or w nodes. I don’t think there is a need to have a pattern the way they are written.

You can use the following if you want to remove all the relationships connected to node ‘n’

Match(n:XYZ{id:100})
Match(n)-[r]-()
Delete r

Or, using your forEach pattern

Match(n:XYZ{id:100})
ForEach(x in [(n)-[r]-()|r] | delete x)
3

If the car's make matches, will it erase the previous labels for Country, Make, and Color? What I actually need is for a node to hold onto labels for Manufacturer, Country, Make, and Color, and then add any specified labels if a match is detected in a FOREACH statement.

For instance, a node with the make "BMW" should have labels for Manufacturer, Country, Make, Color, BMW, and Germany. Similarly, a node with the make "Tesla" should have labels for Manufacturer, Country, Make, Color, Tesla, America, and Black.

neo4j.topic.cypher.kafka_car = 
WITH event
CALL {
	WITH event
	MERGE (n:Manufacturer {id: event.id})
	ON CREATE SET n += event
	ON MATCH SET n += event
	WITH n
	FOREACH ( _ IN CASE WHEN n.Make = 'BMW' THEN [1] ELSE [] END | SET n:Country:Make:Color:Bmw:Germany)
	FOREACH ( _ IN CASE WHEN n.Make = 'Mercedes' THEN [1] ELSE [] END | SET n:Country:Make:Color:Mercedes:Germany)
	FOREACH ( _ IN CASE WHEN n.Make = 'Land Rover' THEN [1] ELSE [] END | SET n:Country:Make:Color:LandRover:England)
	FOREACH ( _ IN CASE WHEN n.Make = 'Tesla' THEN [1] ELSE [] END | SET n:Country:Make:Color:Tesla:America:Black)
	FOREACH ( x IN [(n)-[r]-()|r] | DELETE x)
	FOREACH ( q IN [(q:Manufacturer {ParentName: n.Name})-[*0]-() | q] | MERGE (n)-[:Parent_To]->(q))
	FOREACH ( w IN [(w:Engine {Chasis: n.SerialNumber})-[*0]-() | w] | MERGE (n)-[:Has_Engine]->(w))
}

Will the revised query effectively delete the relationship, ensuring only the latest relationship is retained and multiple node labels is set as my requirement?

4

It should.

Btw, the original Manufacturer label is lost after this query completes. Is that intended?

5

No, it should not. I need "Manufacturer" too. So, I need to add it in every FOREACH for setting additional labels right.

Also, FOREACH ( x IN [(n)-[r]-()|r] | DELETE x) will this ensure delete the old relationships for the particular incoming node?

6

Yes you will. You could use APOC procedures to add/remove individual labels.

Using them will make your code dependent on the APOC library.

There are also procedures to perform conditional logic.

I prefer the simplest approach that uses cypher , so I only use the APOC when necessary.

7

Sounds good. Thanks for your help. Let me try it.