数据存入HBase
FluxInfo代码:
这是一个javabean,用来封装tuple中各字段信息,然后存入hbase中。一条访问记录就是一个实例对象。
有一处需要注意:在插入hbase表时,需要指定行键。我们规定的行键规则为: sstime_uvid_ssid_随机数;
package cn.tarena.domain;
public class FluxInfo {
private String url;
private String urlname;
private String uvid;
private String ssid;
private String sscount;
private String sstime;
private String cip;
public String getRK(){
return sstime+"_"+uvid+"_"+ssid+"_"+(int)(Math.random()*100);
}
public String getUrl() {
return url;
}
public void setUrl(String url) {
this.url = url;
}
public String getUrlname() {
return urlname;
}
public void setUrlname(String urlname) {
this.urlname = urlname;
}
public String getUvid() {
return uvid;
}
public void setUvid(String uvid) {
this.uvid = uvid;
}
public String getSsid() {
return ssid;
}
public void setSsid(String ssid) {
this.ssid = ssid;
}
public String getSscount() {
return sscount;
}
public void setSscount(String sscount) {
this.sscount = sscount;
}
public String getSstime() {
return sstime;
}
public void setSstime(String sstime) {
this.sstime = sstime;
}
public String getCip() {
return cip;
}
public void setCip(String cip) {
this.cip = cip;
}
}
HBaseBolt代码示例:
此类是一个Bolt组件,是最下游的Tuple。用于将数据存入Hbase中。
package cn.tarena.weblog;
import java.util.Map;
import backtype.storm.task.OutputCollector;
import backtype.storm.task.TopologyContext;
import backtype.storm.topology.OutputFieldsDeclarer;
import backtype.storm.topology.base.BaseRichBolt;
import backtype.storm.tuple.Tuple;
import cn.tarena.dao.HBaseDao;
import cn.tarena.domain.FluxInfo;
/*
*
*/
public class HBaseBolt extends BaseRichBolt{
private OutputCollector collector = null;
@Override
public void prepare(Map stormConf, TopologyContext context, OutputCollector collector) {
this.collector = collector;
}
@Override
public void execute(Tuple input) {
try {
FluxInfo fi = new FluxInfo();
fi.setUrl(input.getStringByField("url"));
fi.setUrlname(input.getStringByField("urlname"));
fi.setUvid(input.getStringByField("uvid"));
fi.setSsid(input.getStringByField("ssid"));
fi.setSstime(input.getStringByField("sstime"));
fi.setSscount(input.getStringByField("sscount"));
fi.setCip(input.getStringByField("cip"));
//自定义的dao层工具类,用于存储数据以及其他的dao层操作
HBaseDao.getHbaseDao().saveToHbase(fi);
collector.ack(input);
} catch (Exception e) {
e.printStackTrace();
collector.fail(input);
}
}
@Override
public void declareOutputFields(OutputFieldsDeclarer declarer) {
}
}
HBaseDao 代码示例:
package cn.tarena.dao;
import java.io.IOException;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.hbase.client.HTable;
import org.apache.hadoop.hbase.client.Put;
import cn.tarena.domain.FluxInfo;
public class HBaseDao {
private static HBaseDao hbaseDao = new HBaseDao();
private HBaseDao() {
}
public static HBaseDao getHbaseDao(){
return hbaseDao;
}
/**
* 将信息写入hbase
* @param fi 封装了日志信息的bean
*/
public void saveToHbase(FluxInfo fi){
HTable tab = null;
try {
//1.创建hbase配置对象
Configuration conf = new Configuration();
//--其他用默认配置 至少要配置zookeeper的地址 客户端通过链接zookeeper获取元数据信息
conf.set("hbase.zookeeper.quorum",
"192.168.150.137:2181,192.168.150.138:2181,192.168.150.139:2181");
//2.创建HTable对象
tab = new HTable(conf,"flux".getBytes());
//3.向表中存入数据
Put put = new Put(fi.getRK().getBytes());
put.add("cf1".getBytes(), "url".getBytes(), fi.getUrl().getBytes());
put.add("cf1".getBytes(), "urlname".getBytes(), fi.getUrlname().getBytes());
put.add("cf1".getBytes(), "uvid".getBytes(), fi.getUvid().getBytes());
put.add("cf1".getBytes(), "ssid".getBytes(), fi.getSsid().getBytes());
put.add("cf1".getBytes(), "sscount".getBytes(), fi.getSscount().getBytes());
put.add("cf1".getBytes(), "sstime".getBytes(), fi.getSstime().getBytes());
put.add("cf1".getBytes(), "cip".getBytes(), fi.getCip().getBytes());
tab.put(put);
} catch (IOException e) {
e.printStackTrace();
} finally {
//4.关闭连接
try {
tab.close();
} catch (IOException e) {
e.printStackTrace();
}
}
}
}
注:清空habase表指令: truncate ‘表名’
PvBolt+UvBolt+VvBolt的改造
业务说明
对于pv、uv和vv的统计,我们要实现的效果是:
1)每当一个用户访问一次,storm就接收一条记录
2)pv的处理:让每条记录的pv=1,即并不是马上做叠加处理
3)uv的处理:根据每条记录的uvid,去hbase中去查询,如果没有查到记录,让每条记录的uv=1。如果查到,则让每条记录的uv=0
4)vv的处理:根据每条记录的sscount,如果ssount=0,让每条记录的vv=1,反之vv=0
5)前面4步并没有做指标的叠加统计,所以我们最后会将所有记录都插入到数据库中,如下:
这么做的目的在于我们可以从数据库中,查询一段时间内的总的pv、uv等指标数据。
PvBolt代码示例:
import java.util.List;
import java.util.Map;
import backtype.storm.task.OutputCollector;
import backtype.storm.task.TopologyContext;
import backtype.storm.topology.OutputFieldsDeclarer;
import backtype.storm.topology.base.BaseRichBolt;
import backtype.storm.tuple.Fields;
import backtype.storm.tuple.Tuple;
public class PvBolt extends BaseRichBolt{
private OutputCollector collector;
@Override
public void prepare(Map stormConf, TopologyContext context, OutputCollector collector) {
this.collector=collector;
}
@Override
public void execute(Tuple input) {
try {
List<Object> values=input.getValues();
//--令每条记录的pv=1
values.add(1);
//--锚定
collector.emit(input,values);
collector.ack(input);
} catch (Exception e) {
collector.fail(input);
e.printStackTrace();
}
}
@Override
public void declareOutputFields(OutputFieldsDeclarer declarer) {
declarer.declare(new Fields("url","urlname","uvid","ssid","sscount","sstime","cip","pv"));
}
}
UvBolt代码示例:
package cn.tarena.store.weblog;
import java.util.Calendar;
import java.util.List;
import java.util.Map;
import backtype.storm.task.OutputCollector;
import backtype.storm.task.TopologyContext;
import backtype.storm.topology.OutputFieldsDeclarer;
import backtype.storm.topology.base.BaseRichBolt;
import backtype.storm.tuple.Fields;
import backtype.storm.tuple.Tuple;
import cn.tarena.dao.HBaseDao;
import cn.tarena.domain.FluxInfo;
public class UvBolt extends BaseRichBolt{
private OutputCollector collector;
@Override
public void prepare(Map stormConf, TopologyContext context, OutputCollector collector) {
this.collector=collector;
}
@Override
public void execute(Tuple input) {
try {
//--获取uvid
String uvid = input.getStringByField("uvid");
//--获取sstime
long endTime = Long.parseLong(input.getStringByField("sstime"));
//--基于当前时间向前寻找今天0点的时间值
Calendar calendar = Calendar.getInstance();
calendar.setTimeInMillis(endTime);
calendar.set(Calendar.HOUR, 0);
calendar.set(Calendar.MINUTE,0);
calendar.set(Calendar.SECOND,0);
calendar.set(Calendar.MILLISECOND,0);
long beginTime = calendar.getTimeInMillis();
//--从hbase中查询 今天0点到当前日志时间的数据 中uvid和当前uvid相同的数据
String regex="^\\d{13}_"