#! /bin/bash # # # # # # # # # # # # # # # # # # # # # # # @file :extract_device.sh # @author :liushuai # @revision:2017-01-18 17:01 # # # # # # # # # # # # # # # # # # # # # # source ../dmp_env.sh LOG_TIME=$(date +%Y%m%d -d "-1 day $ScheduleTime") year=${LOG_TIME:0:4} month=${LOG_TIME:4:2} day=${LOG_TIME:6:2} INPUT_PATH="${GENDER_MERGE_INSTALL}/${year}/${month}/${day}/" OUTPUT_PATH="${GENDER_MERGE_DEVICE}/${year}/${month}/${day}/" FILEDICT="${GENDER_CALC_PACKAGE_DICT}/${year}/${month}/${day}/part-00000" RM_LOG_TIME=$(date +%Y/%m/%d -d "-10 day $ScheduleTime") RM_DATE=$(date +%Y%m%d -d "-10 day $ScheduleTime") RM_OUTPUT_PATH="${GENDER_MERGE_DEVICE}/${RM_LOG_TIME}" spark-submit --class mobvista.dmp.datasource.age_gender.CalcDeviceGender \ --name "CalcDeviceGender.${LOG_TIME}" \ --conf spark.sql.shuffle.partitions=10000 \ --conf spark.default.parallelism=2000 \ --conf spark.sql.autoBroadcastJoinThreshold=31457280 \ --conf spark.storage.memoryFraction=0.4 \ --conf spark.shuffle.memoryFraction=0.4 \ --conf spark.executor.extraJavaOptions="-XX:+UseG1GC" \ --conf spark.sql.files.maxPartitionBytes=268435456 \ --deploy-mode cluster --executor-memory 8g --driver-memory 4g --executor-cores 5 --num-executors 100 \ ../${JAR} -date ${LOG_TIME} -merge_input ${INPUT_PATH} -dict_input ${FILEDICT} -output ${OUTPUT_PATH} -parallelism 500 if [[ $? -ne 0 ]]; then exit 255 fi mount_partition "dmp_device_gender" "dt='$LOG_TIME'" "$OUTPUT_PATH" unmount_partition "dmp_device_gender" "dt='$RM_DATE'" "${RM_OUTPUT_PATH}" before_date_path=$(date +%Y/%m/%d -d "-3 day $ScheduleTime") # 删除临时结果 hadoop fs -rm -r ${GENDER_GET_DSP_PATH}/${before_date_path} hadoop fs -rm -r ${GENDER_GET_GA_PATH}/${before_date_path} hadoop fs -rm -r ${GENDER_MERGE_INSTALL}/${before_date_path} hadoop fs -rm -r ${GENDER_CALC_PACKAGE_DICT}/${before_date_path} exit 0