我尝试从 HDFS 中的文件聚合数据。 我需要从那些对 hbase 中的特定表具有值(value)的数据中添加一些详细信息。

但我有异常(exception):

org.apache.spark.SparkException: Task not serializable
    at org.apache.spark.util.ClosureCleaner$.ensureSerializable(ClosureCleaner.scala:166)
    at org.apache.spark.util.ClosureCleaner$.clean(ClosureCleaner.scala:158)
    at org.apache.spark.SparkContext.clean(SparkContext.scala:1623)
    at org.apache.spark.rdd.RDD.map(RDD.scala:286)
    at org.apache.spark.api.java.JavaRDDLike$class.mapToPair(JavaRDDLike.scala:113)
    at org.apache.spark.api.java.AbstractJavaRDDLike.mapToPair(JavaRDDLike.scala:46)
    at ......
    at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
    at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:62)
    at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
    at java.lang.reflect.Method.invoke(Method.java:497)
    at org.apache.spark.deploy.SparkSubmit$.org$apache$spark$deploy$SparkSubmit$$runMain(SparkSubmit.scala:577)
    at org.apache.spark.deploy.SparkSubmit$.doRunMain$1(SparkSubmit.scala:174)
    at org.apache.spark.deploy.SparkSubmit$.submit(SparkSubmit.scala:197)
    at org.apache.spark.deploy.SparkSubmit$.main(SparkSubmit.scala:112)
    at org.apache.spark.deploy.SparkSubmit.main(SparkSubmit.scala)
Caused by: java.io.NotSerializableException: org.apache.hadoop.hbase.client.ConnectionManager$HConnectionImplementation
Serialization stack:

    at org.apache.spark.serializer.SerializationDebugger$.improveException(SerializationDebugger.scala:38)
    at org.apache.spark.serializer.JavaSerializationStream.writeObject(JavaSerializer.scala:47)
    at org.apache.spark.serializer.JavaSerializerInstance.serialize(JavaSerializer.scala:80)
    at org.apache.spark.util.ClosureCleaner$.ensureSerializable(ClosureCleaner.scala:164)

我知道问题发生在我们尝试在 map 函数期间访问 hbase 时。

我的问题是:如何使用 hbase 表中包含的值来完成我的 RDD。

例如: hdfs 中的文件是 csv:

Name;Number1;Number2
toto;1;2

在 hbase 中，我们有与名称 toto 关联的数据。

我需要检索数字 1 和数字 2 的总和(这是最简单的部分) 并与表中的数据聚合。 例如:

reducer 的键将是 tata 并通过获取 hbase 表中的行键 toto 来检索。

有什么建议吗？

最佳答案

最后一位同事做到了，多亏了你的建议:

所以这是允许使用 hbase 表中的数据聚合文件的映射代码。

private final Logger LOGGER = LoggerFactory.getLogger(AbtractGetSDMapFunction.class);




/**
 * Namespace name
 */
public static final String NAMESPACE = "NameSpace";
private static final String ID = "id";
private Connection connection = null;
private static final String LINEID = "l";
private static final String CHANGE_LINE_ID = "clid";
private static final String CHANGE_LINE_DATE = "cld";
private String constClientPortHBase;
private String constQuorumHBase;
private int constTimeOutHBase;
private String constZnodeHBase;
public void initConnection() {
    Configuration conf = HBaseConfiguration.create();
    conf.setInt("timeout", constTimeOutHBase);
    conf.set("hbase.zookeeper.quorum", constQuorumHBase);
    conf.set("hbase.zookeeper.property.clientPort", constClientPortHBase);
    conf.set("zookeeper.znode.parent", constZnodeHBase);
    try {
        connection = HConnectionManager.createConnection(conf);
    } catch (Exception e) {
        LOGGER.error("Error in the configuration of the connection with HBase.", e);
    }
}

 public Tuple2<String, myInput> call(String row) throws Exception {
//this is where you need to init the connection for hbase to avoid serialization problem
    initConnection();

....do your work 
State state = getCurrentState(myInput.getKey());
....do your work 
}

public AbtractGetSDMapFunction( String constClientPortHBase, String constQuorumHBase, String constZnodeHBase, int constTimeOutHBase) {
    this.constClientPortHBase = constClientPortHBase;
    this.constQuorumHBase = constQuorumHBase;
    this.constZnodeHBase = constZnodeHBase;
    this.constTimeOutHBase = constTimeOutHBase;
}

/***************************************************************************/
/**
 * Table Name
 */
public static final String TABLE_NAME = "Table";

public state getCurrentState(String key) throws TechnicalException {
    LOGGER.debug("start key {}", key);
    String buildRowKey = buildRowKey(key);
    State currentState = new State();
    String columnFamily = State.getColumnFamily();
    if (!StringUtils.isEmpty(buildRowKey) && null != columnFamily) {
        try {
            Get scan = new Get(Bytes.toBytes(buildRowKey));
            scan.addFamily(Bytes.toBytes(columnFamily));
            addColumnsToScan(scan, columnFamily, ID);                
            Result result = getTable().get(scan);
            currentState.setCurrentId(getLong(result, columnFamily, ID));              
        } catch (IOException ex) {
            throw new TechnicalException(ex);
        }
        LOGGER.debug("end ");
    }
    return currentState;
}

/***********************************************************/

private Table getTable() throws IOException, TechnicalException {
    Connection connection = getConnection();
    // Table retrieve
    if (connection != null) {
        Table table = connection.getTable(TableName.valueOf(NAMESPACE, TABLE_NAME));


        return table;
    } else {
        throw new TechnicalException("Connection to Hbase not available");
    }
}

/****************************************************************/




private Long getLong(Result result, String columnFamily, String qualifier) {
    Long toLong = null;
    if (null != columnFamily && null != qualifier) {
        byte[] value = result.getValue(Bytes.toBytes(columnFamily), Bytes.toBytes(qualifier));
        toLong = (value != null ? Bytes.toLong(value) : null);
    }
    return toLong;
}

private String getString(Result result, String columnFamily, String qualifier) {
    String toString = null;
    if (null != columnFamily && null != qualifier) {
        byte[] value = result.getValue(Bytes.toBytes(columnFamily), Bytes.toBytes(qualifier));
        toString = (value != null ? Bytes.toString(value) : null);
    }
    return toString;
}


public Connection getConnection() {
    return connection;
}

public void setConnection(Connection connection) {
    this.connection = connection;
}



private void addColumnsToScan(Get scan, String family, String qualifier) {
    if (org.apache.commons.lang.StringUtils.isNotEmpty(family) && org.apache.commons.lang.StringUtils.isNotEmpty(qualifier)) {
        scan.addColumn(Bytes.toBytes(family), Bytes.toBytes(qualifier));
    }
}

private String buildRowKey(String key) throws TechnicalException {
    StringBuilder rowKeyBuilder = new StringBuilder();
    rowKeyBuilder.append(HashFunction.makeSHA1Hash(key));
    return rowKeyBuilder.toString();
}

关于java - 在 map 调用中获取 Spark 上的行，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/41985552/