0
votes

I am having strange issue with my batch inserter. The batch inserter works fine but when I start the server from that location, CYPHER is not able to filter on property.

Query "match(a) return a" returns all nodes. But the moment I try to filter it based on any property then it does not return any row. Query match(a) where a.Name="Someone" return a does not return anything.

If I run run SET command to update the property then I am able to filter it fine. Looks like index issue but not able to figure out exactly.

**Output of - `match(a) return a` -** 

╒══════════════════════════════╕
│a                             │
╞══════════════════════════════╡
│{}                            │
├──────────────────────────────┤

│{Company: "Neo Technology", ye│
│ar: 2013, Name: "Kenny Bastani│
│"}                            │
├──────────────────────────────┤
│{Company: "Neo Technology", ye│
│ar: 2010, Name: "Michael Hunge│
│r"}                           │
├──────────────────────────────┤
│{Company: "Heroku", year: 2011│
│, Name: "James Ward"}         │
├──────────────────────────────┤
│{Name: "Someone"}             │
├──────────────────────────────┤
│{Company: "Doe.com", year: "ni│
│netynine", Name: "John"}      │
└──────
**File -** 

    Name,Company,year

    "Kenny Bastani","Neo Technology",2013

     "Michael Hunger","Neo Technology",2010

     "James Ward","Heroku",2011

     "Someone",,

     "John","Doe.com","ninetynine"

**Batch Inserter-** 

    public void importNodes() throws IOException {
            FileInputStream fis = new FileInputStream(
                    "E:\\neo4j\\CSV_Import\\01.csv");
            InputStreamReader isr = new InputStreamReader(fis);
            BufferedReader br = new BufferedReader(isr);
            String line = null;

            File storeDir = new File("E:\\neo4j\\test_data1");
            FileUtils.deleteRecursively(storeDir);


            BatchInserter neoBatchInserter = BatchInserters.inserter(storeDir);

            Map properties = new HashMap();
            ArrayList headerMap = new ArrayList();

            BatchInserterIndexProvider indexProvider = new LuceneBatchInserterIndexProvider(
                    neoBatchInserter);

            BatchInserterIndex autoIndex = indexProvider.nodeIndex("node_auto_index",
                    MapUtil.stringMap("type", "exact"));

            autoIndex.setCacheCapacity( "Name", 10000 );

            neoBatchInserter.createDeferredSchemaIndex(Labels.Person).on("Name").create();

            String propertyName = "";

            boolean headerFlag = true;

            int counter = 0;

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

                String[] values = line.split(",");

                for (String value : values) {

                    if (headerFlag) {

                        headerMap.add(value);

                    } else {

                        propertyName = headerMap.get(counter);
                        properties.put(propertyName, value);

                    }

                    counter++;
                }

                System.out.println("New Node - " + properties);
                long node = neoBatchInserter.createNode(properties, Labels.Person);
                autoIndex.add(node, properties);
                autoIndex.flush();

                counter = 0;
                properties.clear();
                headerFlag = false;

            }

            indexProvider.shutdown();
            neoBatchInserter.shutdown();
        }
1
The code looks okish, you don't need the node_auto_index - Michael Hunger
You have to remove the '"' or better use a proper CSV parser like OpenCSV or even better use the neo4j-import tool. - Michael Hunger

1 Answers

0
votes

If it is an index issue, deleting and recreating the index may fix it.