CnGoodChannel.scala 17.4 KB
Newer Older
fan.jiang committed
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38
package mobvista.dmp.datasource.dm

import mobvista.dmp.common.CommonSparkJob
import mobvista.dmp.format.RDDMultipleOutputFormat
import org.apache.commons.cli.Options
import org.apache.hadoop.fs.{FileSystem, Path}
import org.apache.hadoop.io.SequenceFile.CompressionType
import org.apache.hadoop.io.Text
import org.apache.hadoop.io.compress.{CompressionCodec, GzipCodec}
import org.apache.spark.rdd.RDD
import org.apache.spark.sql.functions.{concat_ws, lit}
import org.apache.spark.sql.types.{StringType, StructField, StructType}
import org.apache.spark.sql.{DataFrame, Row, SparkSession}
import org.apache.spark.storage.StorageLevel

import java.net.URI
import scala.collection.mutable.ArrayBuffer

/**
 * @author jiangfan
 * @date 2021/9/9 17:42
 */
class CnGoodChannel extends CommonSparkJob with Serializable {
  override protected def buildOptions(): Options = {
    val options = new Options
    options.addOption("coalesce", true, "[must] coalesce")
    options.addOption("old_data_path", true, "[must] old_data_path")
    options.addOption("output1", true, "[must] output1")
    options.addOption("output2", true, "[must] output2")
    options.addOption("output3", true, "[must] output3")
    options.addOption("output4", true, "[must] output4")
    options.addOption("output5", true, "[must] output5")
    options.addOption("package_name1", true, "[must] package_name1")
    options.addOption("package_name2", true, "[must] package_name2")
    options.addOption("package_name3", true, "[must] package_name3")
    options
  }

fan.jiang committed
39 40 41 42 43 44 45 46 47 48
  def buildRes(row:String): Array[Row]={
    var data = new ArrayBuffer[Row]()
    val length = row.split("\t",-1).length
    if (length == 4) {
      val package_name=row.split("\t", -1)(3)
      data += Row(row.split("\t", -1)(0), row.split("\t", -1)(1), package_name.substring(2,package_name.lastIndexOf("\"")))
    }
    data.toArray
  }

fan.jiang committed
49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
  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 coalesce = commandLine.getOptionValue("coalesce")
    val old_data_path = commandLine.getOptionValue("old_data_path")
    val output1 = commandLine.getOptionValue("output1")
    val output2 = commandLine.getOptionValue("output2")
    val output3 = commandLine.getOptionValue("output3")
    val output4 = commandLine.getOptionValue("output4")
    val output5 = commandLine.getOptionValue("output5")
    val package_name1 = commandLine.getOptionValue("package_name1")
    val package_name2 = commandLine.getOptionValue("package_name2")
    val package_name3 = commandLine.getOptionValue("package_name3")

    val spark = SparkSession.builder()
      .appName("CnGoodChannel")
      .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()

    val sc = spark.sparkContext
    import spark.implicits._

    val conf = spark.sparkContext.hadoopConfiguration
    conf.set("mapreduce.output.compress", "true")
    conf.set("mapreduce.output.compression.codec", "org.apache.hadoop.io.compress.GzipCodec")
    conf.setBoolean("mapreduce.output.fileoutputformat.compress", true)
    conf.set("mapreduce.output.fileoutputformat.compress.type", CompressionType.BLOCK.toString)
    conf.setClass("mapreduce.output.fileoutputformat.compress.codec", classOf[GzipCodec], classOf[CompressionCodec])

    FileSystem.get(new URI(s"s3://mob-emr-test"), spark.sparkContext.hadoopConfiguration).delete(new Path(output1), true)
    FileSystem.get(new URI(s"s3://mob-emr-test"), spark.sparkContext.hadoopConfiguration).delete(new Path(output2), true)
    FileSystem.get(new URI(s"s3://mob-emr-test"), spark.sparkContext.hadoopConfiguration).delete(new Path(output3), true)
    FileSystem.get(new URI(s"s3://mob-emr-test"), spark.sparkContext.hadoopConfiguration).delete(new Path(output4), true)
    FileSystem.get(new URI(s"s3://mob-emr-test"), spark.sparkContext.hadoopConfiguration).delete(new Path(output5), true)

    try {
fan.jiang committed
94 95 96
//  存在这种不符合的oiad类型设备,特殊字符^M,导致\t分隔数据异常,所以添加过滤buildRes函数中的条件    fddeee78^Mb6ff-f118-9ddf-ef9ed3ffcac0

      val old_data: RDD[Row] = sc.textFile(old_data_path).flatMap(buildRes(_))
fan.jiang committed
97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147

      val schema: StructType = StructType(Array(
        StructField("device_id", StringType),
        StructField("device_type", StringType),
        StructField("package_name", StringType)
      ))

      val old_data_df = spark.createDataFrame(old_data, schema)
      old_data_df.createOrReplaceTempView("yesterday_data")


      val sql1=
        s"""
           |select device_id,device_type
           |from
           |  (select  split(onedata,"\\\\|")[0] as device_id,split(onedata,"\\\\|")[1] as device_type
           |  from
           |  (select onedata
           |  from
           |  (
           |      select data
           |      from
           |          (
           |           select concat_ws(",",v1,v2,v3,v4) data
           |           from
           |               (
           |                select concat(imei,"|imei") v1 ,concat(oaid,"|oaid") v2,concat(imeimd5,"|imeimd5") v3 ,concat(oaidmd5,"|oaidmd5") v4
           |                from
           |                   (select
           |                      ( case when imei not in ('0','00000000-0000-0000-0000-000000000000','',' ','00000000','00000000000000','000000000000000','0000000000000000') then imei else null end) as imei,
           |                      ( case when ext_oaid not in ('00000000-0000-0000-0000-000000000000','0000000000000000','0','',' ') then ext_oaid else null end ) as oaid,
           |                      md5(case when imei not in ('0','00000000-0000-0000-0000-000000000000','','00000000','00000000000000','000000000000000','0000000000000000') then imei else null end) as imeimd5,
           |                      md5(case when ext_oaid not in ('00000000-0000-0000-0000-000000000000','0000000000000000','0','',' ') then ext_oaid else null end ) as oaidmd5
           |                    FROM
           |                      dwh.ods_adn_trackingnew_request
           |                    where country_code = 'CN' and platform = 'android' and publisher_id in('11386','20284','16227','23652','18721','20781','22193','21522','21915') and app_id not in ('122317')
           |                      and concat(yyyy,'-',mm,'-',dd)=cast(date_sub(CURRENT_DATE, 1) as string)
           |                    )  tmpdata
           |                )  t01
           |           )  t02  where data!='' ) t03
           |  lateral view explode(split(data, ",")) num  as onedata)   t04
           |  union
           |  select device_id,device_type
           |  from
           |      yesterday_data
           |  where  package_name='${package_name1}'  )  t05
           |  group by device_id,device_type
           |""".stripMargin


      val df01: DataFrame = spark.sql(sql1).persist(StorageLevel.MEMORY_AND_DISK_SER)
fan.jiang committed
148
      df01.createOrReplaceTempView("result_data01")
fan.jiang committed
149

fan.jiang committed
150 151 152 153 154 155 156 157 158 159 160 161 162
      val sql11=
        s"""
           |select device_id,device_type
           |from
           |(select
           |    ( case when device_type  in ('imei','oaid') then md5(device_id)  else device_id end) as device_id,
           |    ( case when device_type  in ('imei','oaid') then concat(device_type,"_md5")  when  device_type='imeimd5' then "imei_md5" when  device_type='oaidmd5' then "oaid_md5" end) as device_type
           |from
           |result_data01) t
           |group by device_id,device_type
           |""".stripMargin

      spark.sql(sql11).rdd.coalesce(coalesce.toInt).map(r => {
fan.jiang committed
163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207
        val device_id = r.getAs[String]("device_id")
        val device_type = r.getAs[String]("device_type")
        (new Text(s"${output1}/${device_type}"), new Text(device_id))
      }).saveAsNewAPIHadoopFile(output1, classOf[Text], classOf[Text], classOf[RDDMultipleOutputFormat[_, _]], conf)


      val sql2=
        s"""
           |select device_id,device_type
           |from
           |  (select  split(onedata,"\\\\|")[0] as device_id,split(onedata,"\\\\|")[1] as device_type
           |  from
           |  (select onedata
           |  from
           |  (
           |      select data
           |      from
           |           (
           |           select concat_ws(",",v1,v2,v3,v4) data
           |           from
           |              (
           |               select concat(imei,"|imei") v1 ,concat(oaid,"|oaid") v2,concat(imeimd5,"|imeimd5") v3 ,concat(oaidmd5,"|oaidmd5") v4
           |               from
           |                     (select
           |                        ( case when imei not in ('0','00000000-0000-0000-0000-000000000000','',' ','00000000','00000000000000','000000000000000','0000000000000000') then imei else null end) as imei,
           |                        ( case when ext_oaid not in ('00000000-0000-0000-0000-000000000000','0000000000000000','0','',' ') then ext_oaid else null end ) as oaid,
           |                        md5(case when imei not in ('0','00000000-0000-0000-0000-000000000000','','00000000','00000000000000','000000000000000','0000000000000000') then imei else null end) as imeimd5,
           |                        md5(case when ext_oaid not in ('00000000-0000-0000-0000-000000000000','0000000000000000','0','',' ') then ext_oaid else null end ) as oaidmd5
           |                     FROM
           |                       dwh.ods_adn_trackingnew_request
           |                     where  country_code = 'CN' and platform = 'android' and publisher_id in('18590','21933','13566','19672','26101','21386')
           |                       and concat(yyyy,'-',mm,'-',dd)=cast(date_sub(CURRENT_DATE, 1) as string)
           |                     )  tmpdata
           |               )  t01
           |           )  t02  where data!='' ) t03
           |  lateral view explode(split(data, ",")) num  as onedata)   t04
           |  union
           |  select device_id,device_type
           |  from
           |      yesterday_data
           |  where  package_name='${package_name2}'  )  t05
           |  group by device_id,device_type
           |""".stripMargin

      val df02: DataFrame = spark.sql(sql2).persist(StorageLevel.MEMORY_AND_DISK_SER)
fan.jiang committed
208 209 210 211 212 213 214 215 216 217 218 219 220
      df02.createOrReplaceTempView("result_data02")

      val sql22=
        s"""
           |select device_id,device_type
           |from
           |(select
           |    ( case when device_type  in ('imei','oaid') then md5(device_id)  else device_id end) as device_id,
           |    ( case when device_type  in ('imei','oaid') then concat(device_type,"_md5")  when  device_type='imeimd5' then "imei_md5" when  device_type='oaidmd5' then "oaid_md5" end) as device_type
           |from
           |result_data02) t
           |group by device_id,device_type
           |""".stripMargin
fan.jiang committed
221

fan.jiang committed
222
      spark.sql(sql22).rdd.coalesce(coalesce.toInt).map(r => {
fan.jiang committed
223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
        val device_id = r.getAs[String]("device_id")
        val device_type = r.getAs[String]("device_type")
        (new Text(s"${output2}/${device_type}"), new Text(device_id))
      }).saveAsNewAPIHadoopFile(output2, classOf[Text], classOf[Text], classOf[RDDMultipleOutputFormat[_, _]], conf)


//  第三个sql需求原本是从adn_dsp.log_adn_dsp_request_orc_hour取数据,但这个表计算量大,发现dwh.etl_dsp_request_daily_hours 保存了adn_dsp.log_adn_dsp_request_orc_hour相关数据,读取该表,减小计算量
//
      val sql3=
        s"""
           |select device_id,device_type
           |from
           |  (select  split(onedata,"\\\\|")[0] as device_id,split(onedata,"\\\\|")[1] as device_type
           |  from
           |  (select onedata
           |  from
           |  (
           |  select concat_ws(",",v1,v2,v3,v4) data
           |  from
           |      (
           |        select concat(imei,"|imei") v1 ,concat(oaid,"|oaid") v2,concat(imeimd5,"|imeimd5") v3 ,concat(oaidmd5,"|oaidmd5") v4
           |        from
           |        (select
           |                 split(exitid,',')[4]  as  imei,
           |                 split(exitid,',')[12] as  oaid,
           |                 split(exitid,',')[5]  as  imeimd5,
           |                 split(exitid,',')[13] as  oaidmd5
           |              from
           |              (select
           |                   packagename,exitid
           |               FROM
           |                   dwh.etl_dsp_request_daily_hours
           |               where   dt=replace(cast(date_sub(CURRENT_DATE, 1) as string),'-','')
           |                     and country = 'CN' and platform = 'android' and exchanges in ('iqiyi','wax')
           |               ) tmpdata lateral view explode(split(packagename, "#")) num  as appid
           |               where appid in ('com.qiyi.video','com.sina.weibo','tv.pps.mobile')
           |         )  t01
           |       )  t02
           |     )  t03
           |    lateral view explode(split(data, ",")) num  as onedata where data!='' and split(onedata,"\\\\|")[0]!='')   t04
           |  union
           |  select device_id,device_type
           |  from
           |      yesterday_data
           |  where  package_name='${package_name3}'  )  t05
           |  group by device_id,device_type
           |""".stripMargin

//      val sql3=
//        s"""
//           |select device_id,device_type
//           |from
//           |  (select  split(onedata,"\\\\|")[0] as device_id,split(onedata,"\\\\|")[1] as device_type
//           |  from
//           |  (select onedata
//           |  from
//           |  (
//           |  select concat_ws(",",v1,v2,v3,v4) data
//           |  from
//           |      (
//           |        select concat(imei,"|imei") v1 ,concat(oaid,"|oaid") v2,concat(imeimd5,"|imeimd5") v3 ,concat(oaidmd5,"|oaidmd5") v4
//           |        from
//           |              (select
//           |                split(ext5,',')[4] as  imei,
//           |                split(ext5,',')[12] as  oaid,
//           |                split(ext5,',')[5] as  imeimd5,
//           |                split(ext5,',')[13] as  oaidmd5
//           |               FROM
//           |                 adn_dsp.log_adn_dsp_request_orc_hour
//           |               where countrycode = 'CN' and os = 'android' and exchanges in ('iqiyi','wax') and appid in ('com.qiyi.video','com.sina.weibo','tv.pps.mobile')
//           |                 and concat(yr,'-',mt,'-',dt)=cast(date_sub(CURRENT_DATE, 1) as string)
//           |              )  tmpdata where imei is not null and oaid is not null and imeimd5 is not null and oaidmd5 is not null
//           |        )  t
//           |     )  tt
//           |  lateral view explode(split(data, ",")) num  as onedata)   ttt
//           |  union
//           |  select device_id,device_type
//           |  from
//           |      yesterday_data
//           |  where  package_name='${package_name3}'  )  tttt
//           |  group by device_id,device_type
//           |""".stripMargin

      val df03: DataFrame = spark.sql(sql3).persist(StorageLevel.MEMORY_AND_DISK_SER)
fan.jiang committed
307 308 309 310 311 312 313 314 315 316 317 318 319
      df03.createOrReplaceTempView("result_data03")

      val sql33=
        s"""
           |select device_id,device_type
           |from
           |(select
           |    ( case when device_type  in ('imei','oaid') then md5(device_id)  else device_id end) as device_id,
           |    ( case when device_type  in ('imei','oaid') then concat(device_type,"_md5")  when  device_type='imeimd5' then "imei_md5" when  device_type='oaidmd5' then "oaid_md5" end) as device_type
           |from
           |result_data03) t
           |group by device_id,device_type
           |""".stripMargin
fan.jiang committed
320

fan.jiang committed
321
      spark.sql(sql33).rdd.coalesce(coalesce.toInt).map(r => {
fan.jiang committed
322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
        val device_id = r.getAs[String]("device_id")
        val device_type = r.getAs[String]("device_type")
        (new Text(s"${output3}/${device_type}"), new Text(device_id))
      }).saveAsNewAPIHadoopFile(output3, classOf[Text], classOf[Text], classOf[RDDMultipleOutputFormat[_, _]], conf)

      val data01 = df01.select(concat_ws("\t", df01.col("device_id"),  df01.col("device_type"),  lit("android"),lit("[\""+package_name1+"\"]")))
      val data01_with_country = df01.select(concat_ws("\t", df01.col("device_id"),  df01.col("device_type"),  lit("android"),lit("CN")))

      val data02 = df02.select(concat_ws("\t", df02.col("device_id"),  df02.col("device_type"),  lit("android"),lit("[\""+package_name2+"\"]")))
      val data02_with_country = df02.select(concat_ws("\t", df02.col("device_id"),  df02.col("device_type"),  lit("android"),lit("CN")))

      val data03 = df03.select(concat_ws("\t", df03.col("device_id"),  df03.col("device_type"),  lit("android"),lit("[\""+package_name3+"\"]")))
      val data03_with_country = df03.select(concat_ws("\t", df03.col("device_id"),  df03.col("device_type"),  lit("android"),lit("CN")))

      data01.union(data02).union(data03).coalesce(coalesce.toInt).write.format("text").mode("overwrite").save(output4)
      data01_with_country.union(data02_with_country).union(data03_with_country).coalesce(coalesce.toInt).write.format("text").mode("overwrite").save(output5)


    } finally {
      spark.stop()
    }
    0
  }

}

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