Mysql-＞Hudi-＞Hive

一 准备

1.启动集群 /hive/mysql

start-all.sh

2.启动spark-shell

spark-shell \
  --master yarn \
//--packages org.apache.hudi:hudi-spark3.1-bundle_2.12:0.12.2 \
  --jars /opt/software/hudi-spark3.1-bundle_2.12-0.12.0.jar \
  --conf 'spark.serializer=org.apache.spark.serializer.KryoSerializer' \
  --conf 'spark.sql.extensions=org.apache.spark.sql.hudi.HoodieSparkSessionExtension'

3.导入依赖包

import org.apache.hudi.DataSourceWriteOptions._
import org.apache.hudi.config.HoodieWriteConfig._
import org.apache.spark.sql.functions._
import org.apache.spark.sql.{SaveMode, SparkSession} 

val tableName="demotable"
val hudiBasePath="hdfs://bigdata1:9000//user/hudi/hudi_ods.db/" + tableName

二 查询mysql数据

val DB_URL="jdbc:mysql://bigdata1:3306/ds_db01?allowPublicKeyRetrieval=true&serverTimezone=UTC&useSSL=false"

连接mysql

val df21 = spark.read.format("jdbc")
    .option("driver", "com.mysql.jdbc.Driver")  // mysql驱动程序类名
    .option("url", DB_URL)		  // 连接url
    .option("dbtable", "customer_inf")	  // 要读取的表
    .option("user", "root")		  // 连接账户，需修改为自己的
    .option("password","123456")		  // 连接密码，需修改为自己的
    .option("fetchsize","50")                  // 每轮读取多少行
    .load()

---

 --------------------  转换Transform-增加静态分区列 -------------
import spark.implicits._

查询

println(df21.count())   
df21.show(10) 

三 追加写入Hudi

val df22 = df21.withColumn("etl_date",lit("20220816"))

// val df22 = df21.withColumn("etl_date",date_format(date_sub(current_date(),1),"yyyyMMdd"))
val dfWithTs = df22.withColumn("ts", current_timestamp())

dfWithTs.write.format("hudi")
.mode(SaveMode.Overwrite)
.option("hoodie.insert.shuffle.parallelism","2")//操作并行度为2
.option("hoodie.upsert.shuffle.parallelism","2")
.option(RECORDKEY_FIELD.key(), "customer_inf_id")//记录键的字段名，作为hudi的主键
.option(PARTITIONPATH_FIELD.key(), "etl_date") 
.option(TBL_NAME.key(), tableName)
.save(hudiBasePath)

.option(PRECOMBINE_FIELD.key(), "InPutTime")//预聚合字段名
.option("hoodie.timestamp.field","modified_time")
.option("hoodie.timestamp.field","birthday")
.option("hoodie.timestamp.field","etl_date")
.option("hoodie.timestamp.field","register_time")

查询

val env_data_df=spark.read.format("org.apache.hudi").load(hudiBasePath)

println(env_data_df.count())

env_data_df.show()

四 外接Hive

val sql_create_table =
      s"""
	     |create table hudi_demo.demotable(
         |customer_inf_id int,
         |  customer_id int,
         |  customer_name string ,
         |  identity_card_type tinyint ,
         |  identity_card_no string,
         |  mobile_phone string,
         |  customer_email string ,
         |  gender string ,
         |  customer_point int,
         |  register_time timestamp ,
         |  birthday date ,
         |  customer_level tinyint  ,
         |  customer_money decimal,
         |  modified_time string,
         |  ts timestamp,
         |  etl_date string
         |)
         |using hudi
         |tblproperties(
         |    primaryKey = 'customer_inf_id',
         |    type = 'cow'
         |)
         | options (
         |    hoodie.metadata.enable = 'true'
         | )
         |partitioned by (etl_date)
         |location '$hudiBasePath'
         |""".stripMargin

spark.sql(sql_create_table)

查询分区

hive查询

FAILED: RuntimeException java.lang.ClassNotFoundException: org.apache.hudi.hadoop.HoodieParquetInputFormat

87308 [a3eed69d-1888-48fb-82f7-7254909d770f main] ERROR org.apache.hadoop.hive.ql.Driver - FAILED: RuntimeException java.lang.ClassNotFoundException: org.apache.hudi.hadoop.HoodieParquetInputFormat

报错Hudi集成Hive时的异常解决方法 java.lang.ClassNotFoundException: org.apache.hudi.hadoop.HoodieParquetInputFormat_田昕峣 Richard的博客-CSDN博客

原因

缺少相应的jar包org.apache.hudi.hadoop.HoodieParquetInputFormat

查看hudi的pom文件发现hive版本为2.3.1

重新编译构建