Spring Repository method complains about field not found

Question

I'm using Spring Data Neo4j in my project and I'm having issues with naming conventions for repositories.

This is a simple class containing only one field and the getters/setters

@RelationshipEntity
public class ScoredRelationship
{
  protected Float score;
}

and the class below extends it with other kind of fields

@RelationshipEntity(
    type = RecommenderRelTypes.GOV_CONSUMER_TO_GOV_CONSUMER_SIMILARITY)
public class GovConsumerToGovConsumerSimilarity extends ScoredRelationship
{ // Other fields}

To access the relationship, I'm using the usual repository class

public interface GovConsumerToGovConsumerSimilarityRepository extends
    GraphRepository<GovConsumerToGovConsumerSimilarity>
{

  public Set<GovConsumerToGovConsumerSimilarity> findByScoreGreaterThan(Float value);

  public Set<GovConsumerToGovConsumerSimilarity>
     findByScoreGreaterThanOrderByScoreDesc(Float value);

  public Set<GovConsumerToGovConsumerSimilarity>
     findTopXByScoreGreaterThanOrderByScoreDesc(int limit, Float score);
}

This code compiles well. However, whenever I'm trying to use one of the methods, Spring return a series of exception or doesn't act as intended.

F.e. #findByScoreGreaterThan(0.3f) always returns an empty set. However, invoking a findAll() and printing all the scores it actually has a lot of objects with a score greater than 0.3.

In the second and third case, it always throws an exception saying

Caused by: Unknown identifier `score`.
    at org.neo4j.cypher.internal.symbols.SymbolTable.evaluateType(SymbolTable.scala:60)
    at org.neo4j.cypher.internal.commands.expressions.Identifier.evaluateType(Identifier.scala:51)
    at org.neo4j.cypher.internal.commands.expressions.Expression.assertTypes(Expression.scala:53)
    at org.neo4j.cypher.internal.pipes.SortPipe$$anonfun$assertTypes$1.apply(SortPipe.scala:34)
    at org.neo4j.cypher.internal.pipes.SortPipe$$anonfun$assertTypes$1.apply(SortPipe.scala:33)
    at scala.collection.LinearSeqOptimized$class.foreach(LinearSeqOptimized.scala:59)
    at scala.collection.immutable.List.foreach(List.scala:45)
    at org.neo4j.cypher.internal.pipes.SortPipe.assertTypes(SortPipe.scala:33)
    at org.neo4j.cypher.internal.pipes.PipeWithSource.<init>(PipeWithSource.scala:27)
    at org.neo4j.cypher.internal.pipes.SortPipe.<init>(SortPipe.scala:29)
    at org.neo4j.cypher.internal.executionplan.builders.SortBuilder.apply(SortBuilder.scala:33)
    at org.neo4j.cypher.internal.executionplan.ExecutionPlanImpl.prepareExecutionPlan(ExecutionPlanImpl.scala:49)
    at org.neo4j.cypher.internal.executionplan.ExecutionPlanImpl.<init>(ExecutionPlanImpl.scala:33)
    at org.neo4j.cypher.ExecutionEngine$$anonfun$prepare$1.apply(ExecutionEngine.scala:67)
    at org.neo4j.cypher.ExecutionEngine$$anonfun$prepare$1.apply(ExecutionEngine.scala:67)
    at org.neo4j.cypher.internal.LRUCache.getOrElseUpdate(LRUCache.scala:37)
    at org.neo4j.cypher.ExecutionEngine.prepare(ExecutionEngine.scala:67)
    at org.neo4j.cypher.ExecutionEngine.execute(ExecutionEngine.scala:59)
    at org.neo4j.cypher.ExecutionEngine.execute(ExecutionEngine.scala:63)
    at org.neo4j.cypher.javacompat.ExecutionEngine.execute(ExecutionEngine.java:79)
    at org.springframework.data.neo4j.support.query.CypherQueryEngine.parseAndExecuteQuery(CypherQueryEngine.java:61)

How could be possible? I mean, the class obviously has the score field. Also, executing the simple #findByScoreGreaterThan(float value) doesn't throw any exception, but at the same the latter method always returns an empty set.

EDIT: These are the queries used by Spring. Actually, they seems right

Executing cypher query: START `govConsumerToGovConsumerSimilarity`=node:__types__(className="it.cerict.recommender.persistence.neo4j.GovConsumerToGovConsumerSimilarity") WHERE `govConsumerToGovConsumerSimilarity`.`score`! > {0} RETURN `govConsumerToGovConsumerSimilarity` params {0=0.3}

Executing cypher query: START `govConsumerToGovConsumerSimilarity`=node:__types__(className="it.cerict.recommender.persistence.neo4j.GovConsumerToGovConsumerSimilarity") WHERE `govConsumerToGovConsumerSimilarity`.`score`! > {0} RETURN `govConsumerToGovConsumerSimilarity` ORDER BY score DESC params {0=0.3}

EDIT2: I've also tried to change the score type from Float to float with no further improvements.

Answer 1

This seems to be a bug related to Spring Data Neo4j. Looking to the query executed, it is clear that it searches for nodes, while it should search for relationships.

I changed the #findByScoreGreaterThanOrderByScoreDesc() method using the @Query annotation that specifies the following Cypher query

START `govConsumerToGovConsumerSimilarity`=rel:__rel_types__(className="it.cerict.recommender.persistence.neo4j.GovConsumerToGovConsumerSimilarity") WHERE `govConsumerToGovConsumerSimilarity`.`score`! > 0.3 RETURN `govConsumerToGovConsumerSimilarity` ORDER BY `govConsumerToGovConsumerSimilarity`.`score` DESC;