茄子快传数据分析（一）----数据清理

茄子快传原理

流程图：

数据

“events”: “1473367236143\u00010\u0001connectByQRCode\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000027\u0001\n1473367261933\u00010\u0001AppLaunch\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000028\u0001\n1473367280349\u00010\u0001connectByQRCode\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000029\u0001\n1473367331326\u00010\u0001AppLaunch\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000030\u0001\n1473367353310\u00010\u0001connectByQRCode\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000031\u0001\n1473367387087\u00010\u0001AppLaunch\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000032\u0001\n1473367402167\u00010\u0001connectByQRCode\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000033\u0001\n1473367451994\u00010\u0001AppLaunch\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000034\u0001\n1473367474316\u00010\u0001connectByQRCode\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000035\u0001\n1473367564181\u00010\u0001AppLaunch\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000036\u0001\n1473367589527\u00010\u0001connectByQRCode\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000037\u0001\n1473367610310\u00010\u0001AppLaunch\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000038\u0001\n1473367624647\u00010\u0001connectByQRCode\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000039\u0001\n1473368004298\u00010\u0001AppLaunch\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000040\u0001\n1473368017851\u00010\u0001connectByQRCode\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000041\u0001\n1473369599067\u00010\u0001AppLaunch\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000042\u0001\n1473369622274\u00010\u0001connectByQRCode\u0001\u00010\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u0001\u00011609072239570000043\u0001\n”,

数据解析

使用json在线解析器解析为：
“header”: {
“cid_sn”: “1501004207EE98AA”, sdn码
“mobile_data_type”: “”,
“os_ver”: “9”, 操作系统
“mac”: “88:1f:a1:03:7d:a8”, 物理地址
“resolution”: “2560x1337”, 分辨率
“commit_time”: “1473399829041”, 提交时间
“sdk_ver”: “103”, sdk版本
“device_id_type”: “mac”, 设备类型
“city”: “江门市”, 城市
“android_id”: “”, 安卓设备的安卓id
“device_model”: “MacBookPro11,1”,设备型号
“carrier”: “中国xx”, 运营商
“promotion_channel”: “1”, 推广渠道
“app_ver_name”: “1.7”, app版本号
“imei”: “”, 入网表示
“app_ver_code”: “23”, 公司内部版本码
“pid”: “pid”,
“net_type”: “3”, 网络类型
“device_id”: “m.88:1f:a1:03:7d:a8”, 设备ip
“app_device_id”: “m.88:1f:a1:03:7d:a8”,
“release_channel”: “appstore”, 发布渠道
“country”: “CN”,
“time_zone”: “28800000”, 时区编码
“os_name”: “ios”, 操作系统类型
“manufacture”: “apple”, 生产厂家
“commit_id”: “fde7ee2e48494b24bf3599771d7c2a78”, 事件标示
“app_token”: “XIAONIU_I”, app标示
“account”: “none”, 登陆账号
“app_id”: “com.appid.xiaoniu”, app组名
“build_num”: “YVF6R16303000403”, 编译号
“language”: “zh” 系统所使用语言
}
}

数据清理

1、 数据清理需求分析
release_channel,device_id,city,device_id_type,app_ver_name 这几个字段如果缺失，则过滤
将数据整成 字段,字段,字段,…… 这种形式
在每条数据中添加一个字段：user_id（值就是mac)
2、数据清理代码

public class AppLogClean {
    public static class MapTask extends Mapper<LongWritable, Text, Text, NullWritable> {
        StringBuilder sb = new StringBuilder();
        Text k = new Text();
        @Override
        protected void map(LongWritable key, Text value, Mapper<LongWritable, Text, Text, NullWritable>.Context context)
                throws IOException, InterruptedException {
            // 得到每行数据
            String line = value.toString();
            JSONObject ob1 = JSON.parseObject(line);
            JSONObject ob2 = ob1.getJSONObject("header");



            // 关键数据是否有丢失
            // release_channel,device_id,city,device_id_type,app_ver_name
            // 这几个字段如果缺失，则过滤
            if (StringUtils.isBlank(ob2.getString("release_channel")) 
                    || StringUtils.isBlank(ob2.getString("device_id"))
                    || StringUtils.isBlank(ob2.getString("city"))
                    || StringUtils.isBlank(ob2.getString("device_id_type"))
                    || StringUtils.isBlank(ob2.getString("app_ver_name"))
                    || StringUtils.isBlank(ob2.getString("os_name"))
                    || StringUtils.isBlank(ob2.getString("mac"))) {
                return;
            }

            if (ob2.getString("app_ver_name").equals("android")) {
                if (StringUtils.isBlank(ob2.getString("android_id"))) {
                    return;
                }
            }
            sb.append(ob2.getString("cid_sn")).append(",");
            sb.append(ob2.getString("mobile_data_type")).append(",");
            sb.append(ob2.getString("os_ver")).append(",");
            sb.append(ob2.getString("mac")).append(",");
            sb.append(ob2.getString("resolution")).append(",");
            sb.append(ob2.getString("commit_time")).append(",");
            sb.append(ob2.getString("sdk_ver")).append(",");
            sb.append(ob2.getString("device_id_type")).append(",");
            sb.append(ob2.getString("city")).append(",");
            sb.append(ob2.getString("android_id")).append(",");
            sb.append(ob2.getString("device_model")).append(",");
            sb.append(ob2.getString("carrier")).append(",");
            sb.append(ob2.getString("promotion_channel")).append(",");
            sb.append(ob2.getString("app_ver_name")).append(",");
            sb.append(ob2.getString("imei")).append(",");
            sb.append(ob2.getString("app_ver_code")).append(",");
            sb.append(ob2.getString("pid")).append(",");
            sb.append(ob2.getString("net_type")).append(",");
            sb.append(ob2.getString("device_id")).append(",");
            sb.append(ob2.getString("app_device_id")).append(",");
            sb.append(ob2.getString("release_channel")).append(",");
            sb.append(ob2.getString("country")).append(",");
            sb.append(ob2.getString("time_zone")).append(",");
            sb.append(ob2.getString("os_name")).append(",");
            sb.append(ob2.getString("manufacture")).append(",");
            sb.append(ob2.getString("commit_id")).append(",");
            sb.append(ob2.getString("app_token")).append(",");
            sb.append(ob2.getString("account")).append(",");
            sb.append(ob2.getString("app_id")).append(",");
            sb.append(ob2.getString("build_num")).append(",");
            sb.append(ob2.getString("language")).append(",");

            String uid = ob2.getString("mac");//???
            sb.append(uid);
            k.set(sb.toString());
            context.write(k, NullWritable.get());
            //清除sb的数据
            sb.delete(0, sb.length());
        }
    }

    public static void main(String[] args) throws Exception {
        Configuration conf = new Configuration();
        Job job = Job.getInstance(conf);

        //设置map，设置driver，设置输出类型。。。
        job.setJarByClass(AppLogClean.class);
        job.setMapperClass(MapTask.class);

        job.setOutputKeyClass(Text.class);
        job.setOutputValueClass(NullWritable.class);

        FileInputFormat.addInputPath(job, new Path("D:\\a\\appuserdata\\input\\20170101"));
        FileOutputFormat.setOutputPath(job, new Path("D:\\a\\appuserdata\\out"));
        //不需要reduce 可以设置为0
        job.setNumReduceTasks(0);

        boolean ret = job.waitForCompletion(true);
        //System.exit(ret?0:1);
        System.out.println(ret?"你很优秀":"滚去调bug!");
    }
}

3、数据清理完的文件

4、清理完成的数据格式
1501004207EE98AA,,22,1c:77:f6:78:f5:75,1080x1920,1473396818952,103,mac,江门市,867830021735040,
字段与字段间以逗号隔开