package mobvista.dmp.datasource.fmp import java.net.URI import java.text.SimpleDateFormat import java.util.{Calendar, Date} import mobvista.dmp.common.CommonSparkJob import mobvista.dmp.util.JdbcUtils import org.apache.commons.cli.{BasicParser, Options} import org.apache.hadoop.fs.{FileSystem, Path} import org.apache.hadoop.io.compress.GzipCodec import org.apache.spark.sql.SparkSession /** * @package: mobvista.dmp.datasource.fmp * @author: wangjf * @date: 2019-06-10 * @time: 16:45 * @emial: jinfeng.wang@mobvista.com * @phone: 152-1062-7698 */ class FmpInsightJob extends CommonSparkJob { def commandOptions(): Options = { val options = new Options() options.addOption("data", true, "data") options.addOption("output", true, "output") options.addOption("coalesce", true, "coalesce") options } override protected def run(args: Array[String]): Int = { val parser = new BasicParser() val options = commandOptions() val commandLine = parser.parse(options, args) var data = commandLine.getOptionValue("data") var output = commandLine.getOptionValue("output") val coalesce = commandLine.getOptionValue("coalesce") val spark = SparkSession .builder() .appName("FmpInsightJob") .config("spark.rdd.compress", "true") .config("spark.shuffle.compress", "true") .config("spark.sql.orc.filterPushdown", "true") .config("spark.io.compression.codec", "lz4") .config("spark.io.compression.lz4.blockSize", "64k") .config("spark.sql.autoBroadcastJoinThreshold", "209715200") .config("spark.sql.warehouse.dir", "s3://mob-emr-test/spark-warehouse") .config("spark.serializer", "org.apache.spark.serializer.KryoSerializer") .enableHiveSupport() .getOrCreate() try { data = data.replaceAll("""\\\"""","""\\""").replaceAll("""\\""", """\\\"""") val jobProfileEntity: JobProfileEntity = Constant.parseJsonStringJob(data.replaceAll("#*#", " ")) val job_id = jobProfileEntity.job_id output = output + "/" + job_id FileSystem.get(new URI(s"s3://mob-emr-test"), spark.sparkContext.hadoopConfiguration).delete(new Path(output), true) val dfc = Constant.getDF(spark, jobProfileEntity) dfc._1.rdd.map(r => { r.getAs(0).toString }).coalesce(coalesce.toInt).saveAsTextFile(output, classOf[GzipCodec]) val sdf = new SimpleDateFormat("yyyy-MM-dd HH:mm:ss") val return_data_num = dfc._2 val jdbcUtils = new JdbcUtils val connection = jdbcUtils.getConnection("dmp_fmp") val cal = Calendar.getInstance() cal.setTime(new Date()) cal.add(Calendar.HOUR, 13) val update_time = sdf.format(cal.getTime) val sql = s""" |UPDATE dmp_fmp.job_profile SET return_data_num = ${return_data_num}, path = '${output}', flag = 1, update_time = '${update_time}' WHERE job_id = '${job_id}' """.stripMargin println("更新:==>>" + jdbcUtils.update(sql)) } finally { if (spark != null) { spark.stop() } } 0 } } object FmpInsightJob { def main(args: Array[String]): Unit = { new FmpInsightJob().run(args) } }