Cypher query returning node doesn't work with Spark about neo4j-spark-connector

[jexp]

you could also return properties(n)

What is your use-case for returning the full node data?
We could add some code within the connector that turns nodes / relationships into constructs that Spark can work with, e.g maps or flattened objects, but that would require quite some magic which might be confusing for users.

So perhaps an error message that points out properties(n) or even a map projection like: n { id: id(n), labels: labels(n), .* } as node would make more sense?

[jexp]

Also nodes in Neo4j can have different number (and types) of attributes, so flattening them out into columns automatically wouldn't work really well.