package mobvista.dmp.datasource.age_gender

import java.net.URI

import mobvista.dmp.common.CommonSparkJob
import org.apache.commons.cli.Options
import org.apache.hadoop.fs.{FileSystem, Path}
import org.apache.spark.sql.{Row, SaveMode, SparkSession}


class ThirdPartySourceDaily extends CommonSparkJob {

  override protected def buildOptions(): Options = {
    val options = new Options
    options.addOption("outputtotal", true, "[must] outputtotal")
    options.addOption("coalesce", true, "[must] coalesce")
    options.addOption("today", true, "[must] today")
    options.addOption("yesbef3", true, "[must] yesbef3")
    options
  }

  override protected def run(args: Array[String]): Int = {
    val commandLine = commParser.parse(options, args)
    if (!checkMustOption(commandLine)) {
      printUsage(options)
      return -1
    } else printOptions(commandLine)

    val outputtotal = commandLine.getOptionValue("outputtotal")
    val coalesce = commandLine.getOptionValue("coalesce")
    val today = commandLine.getOptionValue("today")
    val yesbef3 = commandLine.getOptionValue("yesbef3")


    val spark = SparkSession.builder()
      .appName("GenderThirdPartySourceDaily")
      .config("spark.rdd.compress", "true")
      .config("spark.io.compression.codec", "snappy")
      .config("spark.sql.orc.filterPushdown", "true")
      .config("spark.sql.warehouse.dir", "s3://mob-emr-test/spark-warehouse")
      .config("spark.serializer", "org.apache.spark.serializer.KryoSerializer")
      .enableHiveSupport()
      .getOrCreate()

    FileSystem.get(new URI(s"s3://mob-emr-test"), spark.sparkContext.hadoopConfiguration).delete(new Path(outputtotal), true)

    try {

      val sql1 =
        s"""
           | select COALESCE(daily.device_id,total.device_id) device_id,
           |  COALESCE(daily.device_type,total.device_type) device_type,
           |  COALESCE(daily.platform,total.platform) platform,
           |  COALESCE(daily.gender,total.gender) gender
           |  from (select device_id,device_type,platform,case when gender = 'male' then 'm'  when gender = 'female' then 'f' else gender end as gender from dwh.etl_gender_thirdparty_data_daily where dt = '${today}'
           |  group by device_id,device_type,platform,gender ) daily
           |  full join
           |  (select device_id,device_type,platform,case when gender = 'male' then 'm'  when gender = 'female' then 'f' else gender end as gender from dwh.etl_gender_thirdparty_data_total where dt = '${yesbef3}'
           |  group by device_id,device_type,platform,gender ) total
           |  on (daily.device_id = total.device_id and daily.device_type = total.device_type and daily.platform = total.platform )
        """.stripMargin

      spark.sql(sql1)/*.filter(line=>{
        val device_id = line.getAs[String]("device_id")
        StringUtils.isNotBlank(device_id) && Pattern.compile("-").split(device_id, -1).length == 5 && !Pattern.compile("^0*-0*-0*-0*-0*$").matcher(device_id)
          .matches()
      })*/.coalesce(coalesce.toInt)
        .write
        .mode(SaveMode.Overwrite)
        .option("orc.compress", "zlib")
        .orc(outputtotal)

    } finally {
      spark.stop()
    }
    0
  }
}

object ThirdPartySourceDaily {
  def main(args: Array[String]): Unit = {
    new ThirdPartySourceDaily().run(args)
  }
}