Cypher executed from Stored Procedure much slower than raw cypher - stored-procedures

Question

I have a query and pattern which I used to run with different parameter values to identify and create nodes. I wanted to make writing with the query simpler, so I put the query in a stored procedure, compiled the jar, and began my processing.
While easier to call the stored procedure and pass the parameters, the time to execute was MUCH slower (around 10 times slower), and was getting progressively worse as I loaded more and more data into the graph. When I switched back to using the raw queries (and more copy/paste) my time to execute dropped back down.
It feels as if the database is recompiling and/or replanning every time the query in the stored procedure calls is run.
Is there a way to cache the query from the stored procedure?
From what I can tell, my code is identical from inside, and outside the stored procedure. The stored procedure runs, just very very slow when compared to calling the cypher outside of the procedure.
Below is my raw cypher query
with ['register'] as verbs match (e:Entity {type:'PRODUCT', graphId: $graphId})
USING INDEX e:Entity(graphId)
with e, verbs
match (e)-[:REFERS]->(eWord:Word {graphId:$graphId})<-[:OBJ|OBL|NMOD]-(vb:Word {graphId: $graphId})-[]->(notWord:Word {graphId: $graphId})
USING INDEX vb:Word(graphId)
USING INDEX notWord:Word(graphId)
where vb.lemma in verbs
create (event:Event {graphId: $graphId, type: 'registerFail'})
with event, e, vb, notWord
merge (event)-[:TRIGGER]->(vb)
merge (event)-[:TRIGGER]->(notWord)
merge (event)-[:RELATED_PRODUCT]->(e)
with event
match (event)-[:TRIGGER]->(word:Word {graphId: $graphId})-[:COMPOUND|COMPOUND_PRT]->(compWord:Word {graphId: $graphId})
USING INDEX word:Word(graphId)
USING INDEX compWord:Word(graphId)
merge (event)-[:TRIGGER]->(compWord);
Here is the code to my stored procedure
#Procedure(name = "ie.createInabilityCypher", mode = Mode.WRITE)
public void createInabilityFromProduct(#Name("listOfVerbs") List<String> verbs, #Name("inabilityType") String inabilityType, #Name("graphId") String graphId) {
String cypherQuery = "" +
"with $verbsList as verbs " +
"match (e:Entity {type:'PRODUCT', graphId: $graphId}) " +
"USING INDEX e:Entity(graphId) " +
"with e, verbs " +
"match (e)-[:REFERS]->(eWord:Word {graphId:$graphId})<-[:OBJ|OBL|NMOD]-(vb:Word {graphId: '" + graphId +"'})-[]->(notWord:Word {graphId: $graphId}) " +
"USING INDEX vb:Word(graphId) " +
"USING INDEX notWord:Word(graphId) " +
"where vb.lemma in verbs " +
"create (event:Event {graphId: $graphId, type: $inabilityType}) " +
"with event, e, vb, notWord " +
"merge (event)-[:TRIGGER]->(vb) " +
"merge (event)-[:TRIGGER]->(notWord) " +
"merge (event)-[:RELATED_PRODUCT]->(e) " +
"with event " +
"match (event)-[:TRIGGER]->(word:Word {graphId: $graphId})-[:COMPOUND|COMPOUND_PRT]->(compWord:Word {graphId: $graphId}) " +
"USING INDEX word:Word(graphId) " +
"USING INDEX compWord:Word(graphId) " +
"merge (event)-[:TRIGGER]->(compWord)";
Map<String, Object> params = new HashMap<>();
params.put("graphId", graphId);
params.put("verbsList", verbs);
params.put("inabilityType", inabilityType);
tx.execute(cypherQuery, params);

Answer 1

You missed one place where the $graphId parameter should be used. That is why the Cypher code is being "recompiled" every time.
Try replacing this snippet:
(vb:Word {graphId: '" + graphId +"'})
with this:
(vb:Word {graphId: $graphId})