spark读取hbase数据

def main(args: Array[String]): Unit = {
      val hConf = HBaseConfiguration.create();
      hConf.set("hbase.zookeeper.quorum","m6,m7,m8") 
      val tableName = "t_person"
      hConf.set(TableInputFormat.INPUT_TABLE, tableName)      
      val hAdmin = new HBaseAdmin(hConf)
      val conf = new SparkConf()
      conf.set("spark.master", "local")
      conf.set("spark.app.name", "spark demo")
      val sc = new SparkContext(conf);
      val rs = sc.newAPIHadoopRDD(hConf, classOf[TableInputFormat], classOf[ImmutableBytesWritable], classOf[Result])
      rs.foreach(x => {
        println(Bytes.toString(x._2.getRow))
        // 通过列族和列名获取列  
        println(Bytes.toInt(x._2.getValue("base_info".getBytes, "age".getBytes)))
      })
  }

 保存数据到hbase数据库中

def main(args: Array[String]): Unit = {
    val conf = new SparkConf()
    conf.set("spark.master", "local")
    conf.set("spark.app.name", "spark demo")
    // 创建SparkSession对象
    val spark = SparkSession.builder().appName("spark sql").config(conf).getOrCreate();
    // 创建sparkContext对象
    val sc = spark.sparkContext
    
    val hbaseConf = HBaseConfiguration.create()
    val tableName = "t_person"
    hbaseConf.set("hbase.zookeeper.quorum","m6,m7,m8")
    hbaseConf.set(TableInputFormat.INPUT_TABLE, tableName)
    
    // val hbaseAdmin = new HBaseAdmin(hbaseConf)
    val jobConf = new JobConf(hbaseConf, this.getClass)
    jobConf.setOutputFormat(classOf[TableOutputFormat])
    jobConf.set(TableOutputFormat.OUTPUT_TABLE, tableName)
    
    val pairs = sc.parallelize(List(("p_0000010", "12")))
    
    def convert(data : (String, String)) = {
      val p = new Put(Bytes.toBytes(data._1))      
      p.add(Bytes.toBytes("base_info"), Bytes.toBytes("age"), Bytes.toBytes(data._2))     
      (new ImmutableBytesWritable , p)         
    }
     
    // 保存数据到hbase数据库中
    new PairRDDFunctions(pairs.map(convert)).saveAsHadoopDataset(jobConf)
  }

  

    原文作者:spark
    原文地址: https://www.cnblogs.com/heml/p/6148139.html
    本文转自网络文章,转载此文章仅为分享知识,如有侵权,请联系博主进行删除。
点赞