Find people who were near to the user's last location - neo4j

Question

I'm trying to find people who were near to a user's last location in last 24 hours for a simple recommendation engine. My data set is simple for test purposes:
What is wrong with my query? It's returning user Ayse multiple times, but I don't want same person twice or more.
Here is my queries:
var neo4j = require('neo4j-driver').v1;
var driver = neo4j.driver("bolt://localhost", neo4j.auth.basic("neo4j", "neo4j"));
var session = driver.session();
function runQuery(query) {
session.run(query, {}).then(function(result){
console.log("finish")
result.records.forEach(function(record) {
console.log(record._fields)
})
}).catch(console.log);
}
//runQuery("MATCH (n) DETACH DELETE n" )
function newUser(user) {
runQuery("CREATE (:User{name:'"+user+"'})")
}
function newLocation(user, lat, lon) {
var q =
"MATCH (u:User{name:'"+user+"'}) "+
"CREATE (l:Location{lat: "+lat+", lon: "+lon+", created_at: TIMESTAMP()}) "+
"CREATE (u)-[:HAVE_BEEN]->(l) "
runQuery(q)
}
/*
newUser("Ozgur")
newUser("Fatma")
newUser("Ayse")
*/
/*
newLocation("Ozgur", 38.134972, 26.96681)
newLocation("Ozgur", 37.239972, 25.96681)
newLocation("Ozgur", 38.334972, 16.96681)
*/
/*
newLocation("Ayse", 38.294972, 26.76681)
newLocation("Ayse", 37.639972, 25.66681)
newLocation("Ayse", 35.134972, 18.96681)
*/
/*
newLocation("Fatma", 31.114972, 21.76681)
newLocation("Fatma", 31.139972, 21.66681)
*/
function findUserRecommendation(user) {
// get user recommendations
var q =
"MATCH "+
//get user locations for last 24 hours (me_loc)
"(me:User{name:'"+user+"'})-[:HAVE_BEEN]->(me_loc:Location), " +
// strangers (recommendations)
"(l:Location)<-[:HAVE_BEEN]-(stranger:User) " +
"WHERE "+
//except me
"NOT me = stranger "+
// nearest location algorithm based on KM
"AND 2 * 6371 * "+
"asin("+
"sqrt("+
"haversin(radians(me_loc.lat - l.lat)) + "+
"cos(radians(me_loc.lat)) * "+
"cos(radians(l.lat)) * haversin(radians(me_loc.lon - l.lon))"+
")"+
// 500 km
") < 500 " +
// last 24 hours of they location
//"AND l.created_at > (TIMESTAMP() - 604800) "+
// last 24 hours of user's location
//"AND me_loc.created_at > (TIMESTAMP() - 604800)" +
// return every stranger's name and last location once
// but returning Ayse's name and different location more than once
"RETURN DISTINCT stranger.name, l.lat, l.lon"
runQuery(q)
}
findUserRecommendation("Ozgur")

Answer 1

DISTINCT outside of an aggregation just ensures that you won't get duplicate result rows, it doesn't limit you to the first row for each DISTINCT element (it actually applies to all non-aggregated elements in the row as a unit, like using a tuple as a key). You'll have to aggregate the lat/long and then figure out which one of them to keep. Assuming you just want any single lat/long, try this in place of your return line:
WITH stranger, COLLECT(l) AS locs
WITH stranger, HEAD(locs) AS l
RETURN stranger.name, l.lat, l.lon
You can replace the HEAD() call with a REDUCE() if you want to pick out a single lat/long based on particular criteria, like most recent or closest.