Cypher query returns results in browser environment but when the same cypher code is used in py2neo function, nothing is returned

Question

Summary of problem

I have attempted to recreate my functions and graphs in neo4j/py2neo using the same csv files but without the unique IDs in the csv files.

However, when I attempt to call the function I wrote in py2neo to return recommended keywords, the function returns an empty list. The same cypher query py2neo uses gives me the correct recommendations when I put it into the browser.

//Jaccard Similarity keyword recommendation  

MATCH (search_query:Title)-[:HAS_KEYWORDS]->(k:Keyword)  
WITH {item:id(search_query), categories: collect(id(k))} as userData  
WITH collect(userData) as data  
CALL algo.similarity.jaccard(data, {topK: 1, similarityCutoff: 0.1, write:true})  
YIELD nodes, similarityPairs, write, writeRelationshipType, writeProperty, min, max, mean, stdDev, p25, p50, p75, p90, p95, p99, p999, p100  
RETURN nodes, similarityPairs, write, writeRelationshipType, writeProperty, min, max, mean, p95  

//Return similar keywords 

MATCH (p:Title)-[:SIMILAR]->(other),  
  (other)-[:HAS_KEYWORDS]->(keyword)    
WHERE not((p)-[:HAS_KEYWORDS]->(keyword)) and p.title contains "aaa"  
RETURN keyword AS keywords  

I get the expected results with this cypher code when I input it into the browser, but I noticed that a warning pops up when I put the cypher query that searches for [:SIMILAR] :

Index cannot execute wildcard query efficiently

If the performance of this statement using CONTAINS doesn't meet your expectations check out the alternative index-providers, see documentation on index configuration. (index is: index on :Title(title))

My py2neo function:

def jacc_kw_rec(search):  
    q_r = []  
    query2 = "MATCH (p:Title)-[:SIMILAR]->(other), (other)-[:HAS_KEYWORDS]->(keyword) WHERE not((p)-[:HAS_KEYWORDS]->(keyword)) and p.Title contains \"{}\"".format(search)+ " RETURN keyword AS keywords LIMIT 3"  

    result = graph.run(query2).data()

    for r in result:
        r_j = json.dumps(r)
        loaded_r = json.loads(r_j)
        kw_ex = loaded_r['keywords']['Keyword']
        q_r.append(kw_ex)
    return q_r

When I tested the old code, the function would return a list of recommended keywords like ['aaa','bbb','ccc']

Update: After testing my function on my old code, it still returns with an empty list. graph.run(query2).data() is empty.

I'd like to know:

1. Why my graph.run(query2).data returns nothing while there are results in the browser.

Thanks a lot,

Eric

Answer 1

...Ended up being a typo I couldn't find due to neo4j thinking it was an empty match.

Due to

kw_ex = loaded_r['keywords']['Keyword']

After changing it to:

kw_ex = loaded_r['keyword']['Keyword']

it started working again.

I still don't understand why the cursor shows up with nothing while in actuality it found something but just encountered an error it could not resolve.