StackOverflowError when using both incoming and outgoing relationships on the same node due to very deep recursive mapping

[duwenice]

We encountered a problem in Spring Data Neo4j when a domain entity contains both incoming and outgoing relationships, and the query returns a deeply nested / complex graph structure (with the same nodes being connected through multi-level in/out relationships).

In such cases, the DefaultNeo4jEntityConverter recursively resolves the relationships and keeps creating nested entities.
When the returned subgraph is deep enough, the recursive mapping ends up consuming too many stack frames and eventually causes a java.lang.StackOverflowError.

@query("""
MATCH (c:analyzer)
WHERE c.id IN $0
CALL apoc.path.subgraphAll(c, {
relationshipFilter: 'child>',
maxLevel: $1
})
YIELD nodes, relationships
UNWIND nodes AS node
OPTIONAL MATCH (node)-[r:child]->(a:analyzer)
RETURN node, collect(r),collect(a)
""")
List findAllChildren(Set tableIds, int maxLevel);

public class TableNode extends HiBaseNode {

@Relationship(type = Neo4jConstants.RelationShip.CHILD, direction = Relationship.Direction.INCOMING)
private Set<TableNode> parentNodes;

@Relationship(type = Neo4jConstants.RelationShip.CHILD)
private Set<TableNode> childNodes;

}

We observed that after increasing the JVM thread stack size (-Xss) the conversion succeeds, which indicates that the problem is caused by deep recursion during property filling.

[meistermeier]

If I understood you right, the problem is caused by the amount of data, Spring Data Neo4j needs to process, right?
Of what amount of hops are we talking here because it's the first time I see this problem caused by actual data.

When you tell me your used Spring Data Neo4j version, I am happy to have a look if we could change anything to make it not fail with default settings. But my feeling is that we also somehow reached the limit of what a generic mapper could do.

[spring-projects-issues]

If you would like us to look at this issue, please provide the requested information. If the information is not provided within the next 7 days this issue will be closed.