一、编写python脚本产生日志

编写一个python脚本向/home/hadoop/app/data/logs/test.log写入对应的日志

#!/usr/bin/env python
# -*- coding:utf-8 -*-
#@Time  : 2019/4/29 12:50
#@Author: qianliu
#@File  : GenerateImooc.py.py

import time
import random

infos = [
    "116.397026,39.918058",
    "116.389275,39.925818",
    "116.287444,39.810742",
    "116.481707,39.940089",
    "116.410588,39.880172",
    "116.394816,39.91181",
    "116.416002,39.952917"
]

phones = [
    "13111111111","13222222222","13333333333",
    "13444444444","13555555555","13666666666",
    "13777777777","13888888888","13999999999"
]

# 产生一个随机的info
def sample_infos():
    return random.sample(infos,1)[0]

# 产生一个随机的电话
def sample_phone():
    return random.sample(phones,1)[0]

def generate_log(count=3):
    time_str = time.strftime("%Y-%m-%d %H:%M:%S",time.localtime())
    f = open("/home/hadoop/app/data/logs/test.log","a+")

    while count >= 1:
        query_log = "{phone}\t{info}\t[{local_time}]".\
            format( phone = sample_phone(), local_time = time_str ,info = sample_infos())
        print(query_log)
        f.write(query_log+"\n")
        count = count - 1

if __name__ == '__main__':
    generate_log(10)

每次运行python脚本就可以从storm接收到一条条日志信息

二、解析字符串

在LogPoccessBolt类的execute方法中加入解析字符串的代码:

/*
             * 进来的字符串形如：13999999999	116.410588,39.880172	[2019-29-04/29/19 13:29:33]
             * 进行解析
             * */
            String[] splits = value.split("\t");
            String phone = splits[0];
            String[] local = splits[1].split(",");
            String longitude = local[0]; //经度
            String latitude = local[1]; //纬度
            long time = DataUtils.getInstance().getTime(splits[2]); //[2019-29-04/29/19 13:29:33]格式的时间转换成一个时间戳

            System.out.println(phone+"\t"+longitude+"\t"+latitude+"\t"+time);

下面是新建的工具类,用于解析时间：

package com.qianliu.Utils;

import org.apache.logging.log4j.core.util.datetime.FastDateFormat;

public class DataUtils {

    private DataUtils(){}

    private static DataUtils instance;

    public static DataUtils getInstance(){
        if(instance == null){
            instance = new DataUtils();
        }

        return instance;
    }
    
	//将"yyyy-MM-dd HH:mm:ss"格式的字符串格式化为1556522469000这样的时间戳
    FastDateFormat format = FastDateFormat.getInstance("yyyy-MM-dd HH:mm:ss");
    public long getTime(String time) throws Exception{
        return format.parse(time.substring(1,time.length()-1)).getTime();
    }
}

测试是否打通所有程序：