在 Cypher 查询中处理空值

Question

我使用Neo4j java driver 将我的mysql 数据库中的节点添加到Neo4j。

我的问题是记录中的某些列值可以为空。我如何在密码查询中处理它们，因为对于 mysql 表中的空值，它没有创建任何节点。

这是我使用 neo4j java 驱动程序创建节点的代码

public void readObsTableAndCreateNode() throws Exception {
        try {
            Class.forName("com.mysql.jdbc.Driver");
            connect = DriverManager.getConnection("jdbc:mysql://localhost/openmrsnew?user=root&password=qwas");
            statement = connect.createStatement();
            resultSet = statement.executeQuery("select obs_id, person_id, concept_id, encounter_id, order_id, obs_datetime, value_text, date_created, creator, uuid, voided from obs where `obs`.`date_created` > '2018-09-20 00:00:00' and concept_id != 33334002 limit 2");

            List<HashMap<String, Object>> cypherQueriesList = new ArrayList<HashMap<String, Object>>();

            while (resultSet.next()) {

                HashMap<String, Object> cipherQueryParams = new HashMap<String, Object>();
                Integer obsId = Integer.parseInt(resultSet.getString("obs_id"));
                Integer personId = Integer.parseInt(resultSet.getString("person_id"));
                Integer conceptId = Integer.parseInt(resultSet.getString("concept_id"));
                Integer creator = Integer.parseInt(resultSet.getString("creator"));
                Integer voided = Integer.parseInt(resultSet.getString("voided"));
                Integer encounterId = encounterId = Integer.parseInt(resultSet.getString("encounter_id"));
                String date_created = resultSet.getString("date_created");
                String value_text= resultSet.getString("value_text");
                String uuid = resultSet.getString("uuid");           
                String obs_datetime= resultSet.getString("obs_datetime");

                ((HashMap<String, Object>) cipherQueryParams).put("creator", creator);
                ((HashMap<String, Object>) cipherQueryParams).put("valueText", value_text);
                ((HashMap<String, Object>) cipherQueryParams).put("uuid", uuid);
                ((HashMap<String, Object>) cipherQueryParams).put("dateCreated", date_created);
                ((HashMap<String, Object>) cipherQueryParams).put("obsDatetime", obs_datetime); 
                ((HashMap<String, Object>) cipherQueryParams).put("obsId", obsId);
                ((HashMap<String, Object>) cipherQueryParams).put("encounterId", encounterId);
                ((HashMap<String, Object>) cipherQueryParams).put("conceptId", conceptId);
                ((HashMap<String, Object>) cipherQueryParams).put("personId", personId);
                ((HashMap<String, Object>) cipherQueryParams).put("voided", voided);

                cypherQueriesList.add(cipherQueryParams);
            }

            HashMap<String, Object> cypherQuerybatch = new HashMap<String, Object>();
            cypherQuerybatch.put("batch", cypherQueriesList);

            // I am using merge because I dont want duplicated enteries

            String cypherQuery = "UNWIND {batch} as row MERGE (m:temp { obsId: row.obsId ,valueText: row.valueText, uuid: row.uuid, dateCreated: row.dateCreated , obsDatetime: row.obsDatetime })"
                    + " WITH m, row MATCH (c:Concept {conceptId: row.conceptId }) MERGE (m)-[:CONCEPT]->(c) "
                    + " WITH m, row MATCH (e:Encounter {encounterId: row.encounterId }) MERGE (m)-[:ENCOUNTER]->(e) "           
                    + " WITH m, row MATCH (u:Users {userId: row.creator }) MERGE (m)-[:USERS]->(u) "
                    + " WITH m, row MATCH (p:Person {personId: row.personId}) MERGE (m)-[:PERSON]->(p) "
                    + " RETURN m";

            runCypherWithParams(cypherQuery, cypherQuerybatch);

        } catch (Exception e) {
            throw e;
        } finally {
            close();
        }

    }



private void runCypherWithParams(String cypherQuery,  Map<String, Object> parameters) throws Exception{
        System.out.println("Posting Data to Neo4J browser...");
        try {
            connNeo4j = new ConNeo4j("bolt://localhost:7687", "neo4j", "qwas");
            connNeo4j.execQueryBySession(cypherQuery, parameters);                
        } catch (Exception e) {
            e.printStackTrace();
        }finally {

        }

    }



public void execQueryBySession(String query, Map<String, Object> parameters) {
        try {
            Session session = driver.session();
            StatementResult result = session.run(query, parameters);

            System.out.println("Success:- " + result);
        } catch (Exception e) {
            System.out.println("[WARNING] Null Row" + e.getMessage());
        }
    }

在 Neo4j 浏览器控制台中，当我在密码查询下运行时，没有形成节点。

MERGE (m:Temp { obsId: 742241 ,valueText: NULL}) return m;

更新：我能够在 Java 端处理空值。但是对于另一张桌子，我遇到了一些奇怪的行为。

在 Neo4j 浏览器控制台中。我将参数设置为

:param batch: [{deathdateEstimated: 0, birthdateEstimated: 1, creator: 6, birthdate: "2018-01-01", dateCreated: "2018-09-20 04:08:0", gender: "M", personId: 1, voided: 0, uuid: "d7132677-6f7b-aeb8-af5f-850d621b6bbb"}, {deathdateEstimated: 0, birthdateEstimated: 1, creator: 1, birthdate: "2018-01-01", dateCreated: "2018-09-20", gender: "M", personId: 2, voided: 0}]

当我在密码查询下运行时，只会创建一个节点而不是两个。

UNWIND {batch} as row MERGE (m:tempPerson) ON CREATE SET m+=row WITH m,row MATCH (u:Users {userId: row.creator }) MERGE (m)-[:USERS]->(u) return m;

Answer 1

来自Neo4j docs：

虽然您可以执行 MATCH，其中匹配中的节点变量是 null（例如来自失败的 OPTIONAL MATCH），您不能创建或 使用空节点变量合并模式。

您可以使用的另一个选项是 Merge 中提供的 ON CREATE 和 ON MATCH 选项。 但是，此查询取决于您的节点可以使用“obsId”字段唯一标识这一事实。

MERGE (m:Temp { obsId: 742241}) 
ON CREATE SET m.valueText = NULL
ON MATCH SET m.valueText = NULL
return m;

Answer 2

仅在这种情况下检查每个输入值是否为“非空”并将其添加到您的cypherQueryParameter HashMap 来解决您在 Java 方面的问题怎么样？