Bug in Node.getRelationships function?

[Sotiris Beis]

I have the following case study. I want to get the neighbors of the certain node. My function is this one:

public Set<Integer> getNeighborsIds(int nodeId) {

Set<Integer> neighbours = new HashSet<Integer>();

try (Transaction tx = neo4jGraph.beginTx()) {

Node n = nodeIndex.get("nodeId", nodeId).getSingle();

for(Relationship relationship : n.getRelationships(Direction.OUTGOING)) {

Node neighbour = relationship.getEndNode();

String neighbourId = (String)neighbour.getProperty("nodeId");

neighbours.add(Integer.valueOf(neighbourId));

}

tx.success();

tx.close();

}

return neighbours;

}

I know that this node has 255 neighbors but this function return only 100 nodes. When I put this line 

System.out.println(IteratorUtil.count(n.getRelationships(Direction.OUTGOING)));

 4 and 5 line the function returns 255 neighbors. The most strange is that the above function prints out the value 100. Is this a bug or can anyone explain this to me?

Thanks,

Sotiris

[Michael Hunger]

Which version of Neo4j are you using?

You use  a Set which elminates duplicates. You probably have duplicate neighbourId's that are only 100 distinct ones.

And you close the transaction twice. It is an auto-closable resource so you can remove your manual tx.close() line.

Cheers,

Michael

----

(michael)-[:SUPPORTS]->(YOU)-[:USE]->(Neo4j)

Learn Online, Offline or Read a Book (in Deutsch)

We're trading T-shirts for cool Graph Models

--
You received this message because you are subscribed to the Google Groups "Neo4j" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neo4j+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Sotiris Beis]

Which version of Neo4j are you using?

I use neo4j-2.1.0-M01

You use  a Set which elminates duplicates. You probably have duplicate neighbourId's that are only 100 distinct ones.

That was my first thought, but I cheched it. There are no dublicates.
Why do you think the result is different when I use this test line

System.out.println(IteratorUtil.count(n.getRelationships(Direction.OUTGOING)));

before the iteration of the relationships?

You received this message because you are subscribed to a topic in the Google Groups "Neo4j" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/neo4j/stHamJpQSBk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to neo4j+un...@googlegroups.com.

[Michael Hunger]

Thanks for the feedback. Could be related to the changes in the store format for heavily connected nodes.

We'll investigate.

Cheers,

Michael

----

(michael)-[:SUPPORTS]->(YOU)-[:USE]->(Neo4j)

Learn Online, Offline or Read a Book (in Deutsch)

We're trading T-shirts for cool Graph Models

[Sotiris Beis]

Ok, can you suggest me another temporary solution?

[Michael Hunger]

Don't use 2.1-M01 for critical development, it is a milestone / feedback release.

[Mattias Persson]

Yup, it's probably a regression introduced in the recent store format changes. I'd love to track this and be able to reproduce it. How were the relationships added to the node? Distribution of types/directions and also in which order they were added. Could you provide detailed information about that, or provide the database zipped up to me directly (mat...@neotechnology.com) ?

Thanks in advance

Mattias Persson, [mat...@neotechnology.com]
Hacker, Neo Technology
www.neotechnology.com

[Sotiris Beis]

Hi Mattias,

here is some more details:

- The function to create the graph

public void createGraphForMassiveLoad(String dbPath) {

System.out.println("Creating Neo4j Graph Database for massive load . . . .");

Map<String, String> config = new HashMap<String, String>();

config.put("cache_type", "none");

config.put("use_memory_mapped_buffers", "true");

config.put("neostore.nodestore.db.mapped_memory", "200M");

config.put("neostore.relationshipstore.db.mapped_memory", "1000M");

config.put("neostore.propertystore.db.mapped_memory", "250M");

config.put("neostore.propertystore.db.strings.mapped_memory", "250M");

inserter = BatchInserters.inserter(dbPath, config);

indexProvider = new LuceneBatchInserterIndexProvider(inserter);

nodes = indexProvider.nodeIndex("nodes", MapUtil.stringMap("type", "exact"));

}

- The code to load the data

public void createGraph(String datasetDir) {

System.out.println("Loading data in massive mode in Neo4j database . . . .");

inserter.createDeferredSchemaIndex(Neo4jGraphDatabase.NODE_LABEL).on("nodeId").create();

try {

BufferedReader reader = new BufferedReader(new InputStreamReader(new FileInputStream(datasetDir)));

String line;

int lineCounter = 1;

// Map<String, Object> properties;

// IndexHits<Long> cache;

long srcNode, dstNode;

while((line = reader.readLine()) != null) {

if(lineCounter > 4) {

String[] parts = line.split("\t");

srcNode = getOrCreate(parts[0]);

dstNode = getOrCreate(parts[1]);

inserter.createRelationship(srcNode, dstNode, Neo4jGraphDatabase.RelTypes.SIMILAR, null);

}

lineCounter++;

}

reader.close();

} 

catch (IOException e) {

e.printStackTrace();

}

nodes.flush();

}

private long getOrCreate(String value) {

Long id = cache.get(Long.valueOf(value));

if(id == null) {

Map<String, Object> properties = MapUtil.map("nodeId", value);

id = inserter.createNode(properties, Neo4jGraphDatabase.NODE_LABEL);

cache.put(Long.valueOf(value), id);

nodes.add(id, properties);

}

return id;

}

- The dataset I made the tests:

http://snap.stanford.edu/data/email-Enron.html

Unfortunately the database is not available right now, but it's pretty easy to reproduce with the above code.

Thanks,

Sotiris

[Mattias Persson]

Thank you, I was just now able to reproduce this!

[Mattias Persson]

This bug has been found and fixed, so will be included in the next milestone.

[Sotiris Beis]

Thanks Mattias.