• HBase按照TimeStamp删除数据


    #!/bin/bash
    
    #两种时间输入,一种是输入起始日期,另一种是直接输入hbase里面数据的起始时间戳
    if [ $# != 5 ];then
            echo 'usage:sh byTimestampRange.sh table "d:pri_key" d "2018-01-22 17:11:52" "2018-01-22 17:14:53"'
            echo 'usage:sh byTimestampRange.sh table "d:pri_key" t "1532599799357" "1532599799357"'
            exit
    fi
    table=$1
    column=$2
    ttype=$3
    stime=$4
    etime=$5
    
    #get startTimestamp,endTimestamp, hbase里面单元格の时间戳为毫秒
    # 命令data -d "@去除后3位的时间戳" 可以查看hbase里面单元格的时间对应的日期形式
    
    #输入为起始日期,先转化为字符串形式的时间戳,到秒,然后开始时间戳补上000,结束时间戳补上999,因为hbase里面时间戳是到毫秒级的
    if [ $ttype == "d" ];then
            startSec=`date -d "$stime" +%s`
            endSec=`date -d "$etime" +%s`
            sTimestamp=$startSec"000"
            eTimestamp=$endSec"999"
            #sTimestamp=$((startSec*1000+`date "+%N"`/1000000))
            #eTimestamp=$((endSec*1000+`date "+%N"`/1000000))
    #直接使用输入的时间戳
    elif [ $ttype == "t" ];then
            sTimestamp=$stime
            eTimestamp=$etime
    else
            echo "timetype:d or t"
            exit
    fi
    
    #echo $table
    #echo $column
    #echo $ttype
    #echo $sTimestamp
    #echo $eTimestamp
    
    currentTime=`date +%s`
    
    echo "scan,get rowkeys,scan: '$table',{ COLUMNS => '$column',TIMERANGE => [$sTimestamp,$eTimestamp]}"
    #notice:[sTimestamp,eTimestamp)!!!
    echo "scan '$table',{ COLUMNS => '$column',TIMERANGE => [$sTimestamp,$eTimestamp]}" | hbase shell > ./scanresult-$currentTime.txt
    
    #删除scan结果文件前面没用的6行
    sed -i '1,6d' scanresult-$currentTime.txt
    
    #删除最后一个空行
    sed -i '$d' scanresult-$currentTime.txt
    
    #删除scan统计条数的行,现在位置在最后
    sed -i '$d' scanresult-$currentTime.txt
    
    #判断下是否查询结果,没有则直接退出
    
    tmpCount=$(wc -l ./scanresult-$currentTime.txt | awk '{print $1}')
    
    if [ "$tmpCount" -eq 0 ];then
            echo "0 rows deleted"
            rm -rf ./scanresult-$currentTime.txt
            exit
    fi
    
    
    #生成hbase 删除语句
    cat scanresult-$currentTime.txt|awk '{print $1}' | while read rowkey
    do
    echo -e "deleteall '${table}','${rowkey}'" >> ./delete-$currentTime.txt
    done
    
    totalCount=$(wc -l ./delete-$currentTime.txt | awk '{print $1}')
    
    echo "exit" >> ./delete-$currentTime.txt
    
    #执行hbase删除
    hbase shell ./delete-$currentTime.txt
    
    echo "$totalCount records deleted"
    
    rm -rf ./scanresult-$currentTime.txt
    rm -rf ./delete-$currentTime.txt

    参考:https://blog.csdn.net/nyistzp/article/details/76922512

  • 相关阅读:
    K2 BPM介绍(2)
    K2 BPM介绍(1)
    认识BPM
    使用VS Code发布博客
    IIS 使用 HTTP/2
    IIS 8的第一次请求不变慢如何配置
    C# 图片识别技术(支持21种语言,提取图片中的文字)
    第九讲 C#练习题
    c#基础 第八讲
    c#基础 第六讲
  • 原文地址:https://www.cnblogs.com/darange/p/9503874.html
Copyright © 2020-2023  润新知