#!/bin/sh set -x export HADOOP_CONF_DIR=/etc/taihao-apps/hadoop-conf HADOOP=/opt/apps/HADOOP-COMMON/hadoop-common-current/bin/hadoop PREDICT_HOME=/root/zhaohp/recommend-emr-dataprocess/predict origin_data_hdfs_dir=/dw/recommend/model/31_ad_sample_data_v3_auto_test bucket_feature_hdfs_dir=/dw/recommend/model/33_ad_train_data_v3_auto_test # 删除五天之前的预测结果文件 delete_predict_5d_ago() { echo "=========== 开始删除五天前的预测结果文件 $(date "+%Y-%m-%d %H:%M:%d") ===========" tmp_file_name=./files_to_delete.txt # 查询五天前的预测结果文件,并保存到临时文件 find "$PREDICT_HOME" -type f -mtime +5 > "${tmp_file_name}" # 逐行读取临时文件中的路径并删除文件 while IFS= read -r file; do echo "Deleting: $file" # rm -f "$file" done < "${tmp_file_name}" # 删除临时文件 rm -f "${tmp_file_name}" echo "=========== 删除五天前的预测结果文件结束 $(date "+%Y-%m-%d %H:%M:%d") ===========" } # 删除五天之前 HDFS中的原始特征数据 delete_hdfs_origin_data_5d_ago() { FIVE_DAYS_AGO=$(date -d "5 days ago" +%Y-%m-%d) $HADOOP fs -ls $origin_data_hdfs_dir | while read line do echo "${line}" dir=$(echo $line | awk '{print $8}') modified_date=$(echo $line | awk '{print $6}') if [[ "${modified_date}" < "${FIVE_DAYS_AGO}" ]]; then echo "dir=${dir}" echo "modified_date=${modified_date}" fi done } main() { # delete_predict_early_5d delete_hdfs_origin_data_5d_ago } main