Spark 增量抽取 Mysql To Hive
作者:mmseoamin日期:2023-12-18

题目要求:

  1. 抽取ds_db01库中customer_inf的增量数据进入Hive的ods库中表customer_inf。根据ods.user_info表中modified_time作为增量字段,只将新增的数据抽入,字段名称、类型不变,同时添加静态分区,分区字段为etl_date,类型为String,且值为当前日期的前一天日期(分区字段格式为yyyyMMdd)。使用hive cli执行show partitions ods.customer_inf命令;

代码实现: 

package org.example
import org.apache.spark.SparkConf
import org.apache.spark.sql.SparkSession
import java.time.LocalDate
object Demo {
  def main(args: Array[String]): Unit = {
//    创建spark
    val conf = new SparkConf().setMaster("local[*]").setAppName("one")
      .set("spark.testing.memory", "2147480000").set("dfs.client.use.datanode.hostname", "true")
    System.setProperty("HADOOP_USER_NAME", "root")
// 连接hive
val spark = SparkSession.builder()
  // 配置 Hive Metastore 的连接地址
  .config("hive.metastore.uris", "thrift://192.168.23.60:9083")
  // 配置 Hive 数据仓库的存储位置
  .config("hive.metastore.warehouse", "hdfs://192.168.23.60://9000/user/hive/warehouse")
  // 配置 Spark SQL 的存储分配策略为 "LEGACY"
  .config("spark.sql.storeAssignmentPolicy", "LEGACY")
  // 添加其他自定义的 Spark 配置
  .config(conf)
  // 启用对 Hive 的支持,使得可以使用 Hive 的表和查询
  .enableHiveSupport()
  // 创建 SparkSession 对象
  .getOrCreate()
//连接mysql
    spark.read.format("jdbc")
      .option("url","jdbc:mysql://192.168.23.60:3306/ds_db01??characterEncoding=UTF-8")
      .option("driver","com.mysql.jdbc.Driver")
      .option("user","root")
      .option("password","123456")
      .option("dbtable","customer_inf")
      .load().createOrReplaceTempView("v")  //对该表创建视图
    spark.sql("select * from v")
//    获取当天时间的前一天
val unit = java.time.LocalDate.now().plusYears(-1).plusMonths(-1).plusDays(-1).toString().replace("-", "")
    val unit1 = unit.toInt
//全量抽取
//    spark.sql(
//      s"""
//         |insert overwrite table gh_test.customer_inf
//         |partition (etl_date="${unit}")
//         |select * from v
//         |
//         |""".stripMargin).show()
//
//spark.sql("select * from gh_test.customer_inf").show
//将modified_time类型转换为yyyyMMdd
    spark.sql(
        s"""
         |select  customer_inf_id,customer_id,customer_name,identity_card_type,identity_card_no,mobile_phone,
         |customer_email,gender,customer_point,register_time,birthday,customer_level,customer_money,
         |from_unixtime(unix_timestamp(modified_time,'yyyy-MM-dd'),'yyyyMMdd') as modified_time
         |from v
         |""".stripMargin).createOrReplaceTempView("v1")
//      spark.sql("select  count(*) from gh_test.customer_inf").show
//从mysql中增量抽取到hive
    spark.sql(
      s"""
         |insert overwrite table gh_test.customer_inf
         |partition (etl_date="${unit}")
         |select * from v where  modified_time>"${unit1}"
         |""".stripMargin
    ).show()
//  spark.sql("select * from gh_test.customer_inf").show
//    查询抽取后的条数
    spark.sql("select  count(*) from gh_test.customer_inf").show
//    spark.sql("desc gh_test.customer_inf")
  }
}