#!/usr/bin/env bash source ../dmp_env.sh source ././../ga_rawdata_analysis/common/tools.sh echo "job begin!!!" dt_today=$(date -d "$ScheduleTime" +"%Y%m%d") dt_slash_today=$(date -d "$ScheduleTime" +"%Y/%m/%d") update=$(date -d "$ScheduleTime" +"%Y-%m-%d") dt_three_days_ago=$(date -d "$ScheduleTime 3 days ago" +"%Y%m%d") INPUT="s3://mob-emr-test/dataplatform/DataWareHouse/data/dev/iqiyi_lahuo_tmp_daily/${dt_slash_today}" #OUTPUT="s3://mob-emr-test/dataplatform/DataWareHouse/data/dwh/iqiyi_lahuo_tmp_daily_to_s3/${dt_today}" OUTPUT="${IQiYi_LAHUO_TMP_DAILY_TO_S3}/${dt_slash_today}/iqiyi_activation" check_await "${INPUT}/_SUCCESS" hadoop fs -rm -r "$OUTPUT" spark-submit --class mobvista.dmp.datasource.iqiyi.IQiYiTmpDataToDMP \ --conf spark.network.timeout=720s \ --conf spark.default.parallelism=2000 \ --conf spark.sql.shuffle.partitions=2000 \ --conf spark.sql.broadcastTimeout=1200 \ --conf spark.yarn.executor.memoryOverhead=4096 \ --conf spark.sql.autoBroadcastJoinThreshold=31457280 \ --master yarn --deploy-mode cluster --executor-memory 8g --driver-memory 4g --executor-cores 4 --num-executors 30 \ ../${JAR} -input ${INPUT} \ -output ${OUTPUT} \ -update ${update} if [ $? -ne 0 ];then exit 255 fi DMP_INSTALL_LIST_OUTPUT_PATH="${DM_INSTALL_LIST}_v2/${dt_slash_today}/iqiyi_activation" mount_partition "dm_install_list_v2" "dt='${dt_today}', business='iqiyi_activation'" "${DMP_INSTALL_LIST_OUTPUT_PATH}" mount_partition "iqiyi_lahuo_tmp_daily_to_s3" "dt='${dt_today}', business='iqiyi_activation'" "${OUTPUT}" FOUR_DAYS_OUTPUT="${IQiYi_LAHUO_TMP_DAILY_TO_S3}/${dt_slash_today}/iqiyi_activation_four_days" spark-submit --class mobvista.dmp.datasource.iqiyi.IQiYiLaHuoFourDaysDataDeduplication \ --conf spark.network.timeout=720s \ --conf spark.default.parallelism=2000 \ --conf spark.sql.shuffle.partitions=2000 \ --conf spark.sql.broadcastTimeout=1200 \ --conf spark.yarn.executor.memoryOverhead=4096 \ --conf spark.sql.autoBroadcastJoinThreshold=31457280 \ --master yarn --deploy-mode cluster --executor-memory 8g --driver-memory 4g --executor-cores 4 --num-executors 40 \ ../${JAR} -dt_today ${dt_today} -dt_three_days_ago ${dt_three_days_ago} \ -output ${FOUR_DAYS_OUTPUT} if [ $? -ne 0 ];then exit 255 fi mount_partition "iqiyi_lahuo_tmp_daily_to_s3" "dt='${dt_today}', business='iqiyi_activation_four_days'" "${FOUR_DAYS_OUTPUT}" if [ $? -ne 0 ];then exit 255 fi HIVE_CMD=$(hive_func) $HIVE_CMD -v -hivevar dt_today ${dt_today} -f iqiyi_tmp_daily_data_to_dmp.sql if [ $? -ne 0 ];then exit 255 fi #hadoop fs -touchz ${DMP_INSTALL_LIST_OUTPUT_PATH}/_SUCCESS echo "good job"