shell按月份循环执行hive脚本

布满荆棘的人生 2021-09-23 22:52 544阅读 0赞

在开发中，怎样编写脚本输入开始月份、结束月份自动跑数呢？

对shell中getopts不懂的请查看我以前博客  
[shell中的getopt][shell_getopt]

提供以下脚本供参考：

#!/bin/sh
    #*****************************************************************************************
    #*** 程序功能: 按天循环导入分区数据
    #*** 输入参数: -s <START_MONTH> :需要导入的开始月份，格式YYYYMM
    #**** -e <END_MONTH> :需要导入的结束月份，格式YYYYMM
    #**** -f <SOL_FILE>:SQL脚本绝对路径名称
    #*** 编 写 人: fuyun
    #*** 编写日期: 2020-02-18
    #*** 修 改 人: 
    #*****************************************************************************************
    
    #若指令传回值不等于0，则立即退出shell
    #set -e
    umask 002
    
    
    while getopts s:e:f: OPT; do #选项后面的冒号表示该选项需要参数
      case ${OPT} in
        s) START_MONTH=${OPTARG}
           ;;
        e) END_MONTH=${OPTARG}
           ;;
        f) SOL_FILE=${OPTARG}
           ;;
        \?)
           printf "[Usage] `date '+%F %T'` `basename $0` -s <START_MONTH:YYYYMM> -e <END_MONTH:YYYYMM> -f <SOL_FILE>\n" >&2
           exit 1
      esac
    done
    
    # check parameter
    if [ -z "${START_MONTH}" -o -z "${END_MONTH}" -o -z "${SOL_FILE}" ]; then
        printf "[ERROR] `date '+%F %T'` following parameters is empty:\nSTART_MONTH=${START_MONTH}\nEND_MONTH=${END_MONTH}\n"
        printf "SOL_FILE=${SOL_FILE}\n"
        exit 1
    fi
    
    LOG_FILE=$(echo "${SOL_FILE}" | awk -F"/" '{print $NF}' | awk -F"." '{print $1}')
    MONTHS=`seq $START_MONTH $END_MONTH`
    
    printf "[INFO] `date '+%F %T'` execute start day is ${START_MONTH}\n"
    printf "[INFO] `date '+%F %T'` execute end day is ${END_MONTH}\n"
    printf "[INFO] `date '+%F %T'` execute SQL file is ${SOL_FILE}\n"
    printf "[INFO] `date '+%F %T'` looping execute.......\n"
    
    executeHive() { 
      for STAT_MONTH in $MONTHS
      do  
        if [ ${STAT_MONTH: 4: 2} -le "12" ] && [ ${STAT_MONTH: 4: 2} -ne "00" ]; then
          local LOG_PATH=/home/fuyun/logs/${LOG_FILE}.log
          printf "[INFO] `date '+%F %T'` log file name is ${LOG_PATH}\n"
          printf "[INFO] `date '+%F %T'` current execute partition is ${STAT_MONTH}\n"
          local startTime=`date '+%F %T'`
          local startSeconds=$(date --date="${startTime}" +%s)
          printf "[INFO] `date '+%F %T'` hivesql execute start time is ${startTime}\n"
          
          # execute sql
          hive --hiveconf MONTH=${STAT_MONTH} -f ${SOL_FILE}>${LOG_PATH}.tmp 2>${LOG_PATH}
          
          exitCode=$?
          if [ $exitCode -ne 0 ];then
            printf "[ERROR] `date '+%F %T'` hivesql execute ${SOL_FILE} is failed!!!\n"
            exit $exitCode
          else
            local endTime=`date '+%F %T'`
            local endSeconds=$(date --date="${endTime}" +%s)
            printf "[INFO] `date '+%F %T'` hivesql execute end time is ${endTime}\n"
            printf "[INFO] `date '+%F %T'` hivesql execute time is $(($((endSeconds-startSeconds))/60)) minutes\n"
            printf "[INFO] `date '+%F %T'` hivesql ${STAT_MONTH} data execute success！！！\n"
            printf "[INFO] `date '+%F %T'` delete temp log ${LOG_PATH}.tmp\n"
            rm ${LOG_PATH}.tmp
          fi
        fi
      done 
    }
    
    if [ ${START_MONTH} -gt ${END_MONTH} ]; then
      printf "[ERROR] `date '+%F %T'` -s:${START_MONTH} greater than -e:${END_MONTH}\n"
    else
      executeHive
      printf "[INFO] `date '+%F %T'` hivesql ${START_MONTH} to ${END_MONTH} data execute success！！！\n"
    fi

[shell_getopt]: https://blog.csdn.net/lz6363/article/details/104190125