自定义聚合函数(统计每种行为的触发次数排名前三的商品id)

package SparkSQL.fun.project

import org.apache.spark.SparkConf
import org.apache.spark.sql.expressions.{MutableAggregationBuffer, UserDefinedAggregateFunction}
import org.apache.spark.sql.types.{DataType, DataTypes, StructField, StructType}
import org.apache.spark.sql.{DataFrame, Dataset, Row, SparkSession}

/**
 * 统计每种行为的触发次数排名前三的商品id
 */
object BehaviorCode2 {
  def main(args: Array[String]): Unit = {
    val sparkConf = new SparkConf().setAppName("project01").setMaster("local[*]")
    val session = SparkSession.builder().config(sparkConf).getOrCreate()
    val map = Map("mode"->"dropMalformed","inferSchema"->"true")
    val frame = session.read.options(map).csv("G:\\shixunworkspace\\sparkcode\\src\\main\\java\\SparkSQL\\fun\\project\\b.csv")

    // "userId", "goodsId", "categoryId", "behavior", "time"

    import session.implicits._
    val frame1: Dataset[UserBehaviorBean] = frame.map(row => {
      UserBehaviorBean(row.getInt(0), row.getInt(1),
        row.getInt(2), row.getString(3), row.getInt(4))
    })
    val frame3 = frame1.toDF("userId", "goodsId", "categoryId", "behavior", "time")
    frame3.createTempView("tmp")

    val frame2 = session.sql("select behavior, goodsId, count(*) count from tmp group by behavior, goodsId")
    frame2.show()

    frame2.createTempView("tmp1")
    val frame4 = session.sql("select behavior, goodsId, count, row_number() over(partition by behavior, goodsId order by count) rn from tmp1")
    frame4.show()

    frame4.createTempView("temp2")
    val frame5 = session.sql("select behavior, goodsId, count, rn from temp2 where rn <= 3")
    frame5.show()

    session.stop()
  }
}

标签：每种,自定义,val,goodsId,session,behavior,sql,id,row
来源： https://www.cnblogs.com/jsqup/p/16659672.html