spark udaf 操作说明

/**
  *  udaf  计算每个评价流水的每个问卷的平均得分
  */
class CalScoreAvl extends UserDefinedAggregateFunction{

  //输入字段的个数和类型 index和score的名称不代表实际的字段 只是一个占位说明
  override def inputSchema: StructType = StructType(Array(StructField("index",StringType),StructField("score",DoubleType)))

  //中间存储的变量定义
  //这里定义三个
  // sum 用来存储总分
  // count 用来存储个数
  // parse 用来存储index的字符串 判断是否重复

  override def bufferSchema: StructType = StructType(Array(StructField("sum",DoubleType),StructField("count",IntegerType),StructField("parse",StringType)))

  //udaf最终的返回值
  override def dataType: DataType = DoubleType

  //默认true就OK
  override def deterministic: Boolean = true

  //初始化 上面定义的三个中间结构变量
  override def initialize(buffer: MutableAggregationBuffer): Unit = {
    buffer.update(0,0.0)
    buffer.update(1,0)
    buffer.update(2,"")
  }

  // 逻辑模块
  // 主要写三个变量通过什么方式去计算的逻辑
  // 这里是字段传入的index没有出现过 而且 score不是 0的话 sum 累加  count + 1  parse的index字符串拼接
  override def update(buffer: MutableAggregationBuffer, input: Row): Unit = {

    if (!buffer.getString(2).contains(input.getString(0)) && input.getDouble(1) != 0){
      buffer(0) = buffer.getDouble(0) + input.getDouble(1)
      buffer(1) = buffer.getInt(1) + 1
      buffer(2) = buffer.getString(2) + "_" + input.getString(0)
    }
  }

  //分布式合并规则
  override def merge(buffer1: MutableAggregationBuffer, buffer2: Row): Unit = {
    buffer1(0) = buffer1.getDouble(0) + buffer2.getDouble(0)
    buffer1(1) = buffer1.getInt(1) + buffer2.getInt(1)
    buffer1(2) = buffer1.getString(2) +""+ buffer2.getString(2)
  }

  //最终输出的结果计算方式
  override def evaluate(buffer: Row): Any = {
    buffer.getDouble(0) / buffer.getInt(1)
  }
}