spark dataframe 数据 转化为 json 或者自定义格式的字符串

  import org.apache.spark.sql.types._
      val schema = StructType(List(
        StructField("integer_column", IntegerType, nullable = true),
        StructField("string_column", StringType, nullable = true),
        StructField("date_column", DateType, nullable = true)
      ))


      val rdd = spark.sparkContext.parallelize(Seq(
        Row(1, "First Value", java.sql.Date.valueOf("2010-01-01")),
        Row(2, "Second Value", java.sql.Date.valueOf("2010-02-01")),
        Row(null, "Second Value", java.sql.Date.valueOf("2010-02-01"))
      ))

df_fill.toJSON.collectAsList.toString

package utils

import org.apache.spark.sql.DataFrame
object MyDataInsightUtil {

  def dataFrame2Json(data:DataFrame,num:Int=10)={
    val dftopN = data.limit(num)

    val arr = dftopN.collect().map(x=>x.toSeq.mkString("\"","\",\"","\"")).mkString("|")

    val columnName = (dftopN.columns.mkString("\"","\",\"","\""))

    (columnName+"|"+ arr)
  }

}

 填充后结果
 “integer_column”,“string_column”,“date_column”|“1”,“First Value”,“2010-01-01”|“2”,“Second Value”,“2010-02-01”|"-3",“Second Value”,“2010-02-01”
 原始结果
 “integer_column”,“string_column”,“date_column”|“1”,“First Value”,“2010-01-01”|“2”,“Second Value”,“2010-02-01”|“null”,“Second Value”,“2010-02-01”