Nutch 2.0 之 Apache Gora MR介绍
-----------------
1. 介绍
Apapche Gora内建了对于Apache Hadoop的支持,而Gora的dataStore可以用来做为InputFormat与OutputFormat的输入与输出,然而这些输出的对象都会被序列化,Gora扩展了Avro的DatumWriters来实现的。
2. 一个简单的例子
下面是gora-tutorial中的一个简单的例子来说明Gora MR的用法,这个例子是使用了上一个例子的输出,作为其输入,用MR来分析日志的信息,用于统计单个URL一天之中的访问量,然后把分析后的结果写回HBase中的一个例子,当然,你也可以把输出写回SQL或者虽的数据库中,这个例子可以在t gora-tutorial/src/main/java/org/apache/gora/tutorial/log/LogAnalytics.java找到。
为了对日志数据进行分析,mapper方法以pageview为输入,输出的Key为<URL,timestamp>对,Value为1,表示一次,在reducer中对相同key的value进行聚合,得到一个总数,然后生成一个MetricDatum对象,把这个对象序列化到数据库中去。
数据模型的定义,这里我们用到了MetricDatum这个对象,它的定义在gora-tutorial/src/main/avro/metricdatum.json,我们要用bin/gora compile来产生它的一个java文件,
json的定义文件如下:
{
"type": "record",
"name": "MetricDatum",
"namespace": "org.apache.gora.tutorial.log.generated",
"fields" : [
{"name": "metricDimension", "type": "string"},
{"name": "timestamp", "type": "long"},
{"name": "metric", "type" : "long"}
]
}
我们可以看到,它的类型为record, 名称空间为org.apache.gora.tutorial.log.generated,还有三个字段,分别定义了字段名与类型。
下面我们还要定义datastore的一个映射文件,内容如下:
<class name="org.apache.gora.tutorial.log.generated.MetricDatum" keyClass="java.lang.String" table="Metrics">
<field name="metricDimension" family="common" qualifier="metricDimension"/>
<field name="timestamp" family="common" qualifier="ts"/>
<field name="metric" family="common" qualifier="metric"/>
</class>
我们可以看到,在class中定义了类名,key的类型,还有表名Metrics。在field中定义了相应的字段名,family还有qualifier。
好了,下面我们来命令这个MR程式,输入如下命令:
$ bin/gora loganalytics
有一些输入,部分如下:
12/07/21 13:35:16 INFO mapreduce.GoraRecordWriter: gora.buffer.write.limit = 10000
12/07/21 13:35:17 INFO mapred.TaskRunner: Task:attempt_local_0001_r_000000_0 is done. And is in the process of commiting
12/07/21 13:35:17 INFO mapred.LocalJobRunner: reduce > reduce
12/07/21 13:35:17 INFO mapred.TaskRunner: Task 'attempt_local_0001_r_000000_0' done.
12/07/21 13:35:17 INFO mapred.JobClient: map 100% reduce 100%
12/07/21 13:35:17 INFO mapred.JobClient: Job complete: job_local_0001
12/07/21 13:35:17 INFO mapred.JobClient: Counters: 12
12/07/21 13:35:17 INFO mapred.JobClient: FileSystemCounters
12/07/21 13:35:17 INFO mapred.JobClient: FILE_BYTES_READ=561645
12/07/21 13:35:17 INFO mapred.JobClient: FILE_BYTES_WRITTEN=1126606
12/07/21 13:35:17 INFO mapred.JobClient: Map-Reduce Framework
12/07/21 13:35:17 INFO mapred.JobClient: Reduce input groups=3033
12/07/21 13:35:17 INFO mapred.JobClient: Combine output records=0
12/07/21 13:35:17 INFO mapred.JobClient: Map input records=10000
12/07/21 13:35:17 INFO mapred.JobClient: Reduce shuffle bytes=0
12/07/21 13:35:17 INFO mapred.JobClient: Reduce output records=3033
12/07/21 13:35:17 INFO mapred.JobClient: Spilled Records=20000
12/07/21 13:35:17 INFO mapred.JobClient: Map output bytes=512510
12/07/21 13:35:17 INFO mapred.JobClient: Combine input records=0
12/07/21 13:35:17 INFO mapred.JobClient: Map output records=10000
12/07/21 13:35:17 INFO mapred.JobClient: Reduce input records=10000
12/07/21 13:35:17 INFO log.LogAnalytics: Log completed with success
完成以后可以查看一下HBase的Metrics表,使用如下命令
hbase(main):007:0> scan 'Metrics' ,{LIMIT=>1}
ROW COLUMN+CELL
/?a=1__-znawtuabsy&k=96804_1236 column=common:metric, timestamp=1342848916670, value=\x00\x00\x00\x00\x00\x00\x00\x09
902400000
/?a=1__-znawtuabsy&k=96804_1236 column=common:metricDimension, timestamp=1342848916670, value=/?a=1__-znawtuabsy&k=96804
902400000
/?a=1__-znawtuabsy&k=96804_1236 column=common:ts, timestamp=1342848916670, value=\x00\x00\x01\x1F\xFD \xD0\x00
902400000
1 row(s) in 0.0200 seconds
我们已经可以看到分析后的数据已经存储到Metrics表中了。
3. 程式代码分析
在定义Hadoop的任务的时候,我们可以选择是否使用Gora来做为其输入与输出, Gora定义了自己的GoraInputFormat 与GoraOutputFormat,而且Gora还定义自己的GoraMapper与GoraReducer类,提供一些方法的初始化静态方法 。在下面的代码中,就使用了Gora(Mapper|reducer)的一些静态方法来进行初始始化。下面是Job初始化代码:
public Job createJob(DataStore<Long, Pageview> inStore
, DataStore<String, MetricDatum> outStore, int numReducer) throws IOException {
Job job = new Job(getConf());
job.setJobName("Log Analytics");
job.setNumReduceTasks(numReducer);
job.setJarByClass(getClass());
/* Mappers are initialized with GoraMapper.initMapper() or
* GoraInputFormat.setInput()*/
GoraMapper.initMapperJob(job, inStore, TextLong.class, LongWritable.class
, LogAnalyticsMapper.class, true);
/* Reducers are initialized with GoraReducer#initReducer().
* If the output is not to be persisted via Gora, any reducer
* can be used instead. */
GoraReducer.initReducerJob(job, outStore, LogAnalyticsReducer.class);
return job;
}
下面是它的run函数的代码:
@Override
public int run(String[] args) throws Exception {
DataStore<Long, Pageview> inStore;
DataStore<String, MetricDatum> outStore;
Configuration conf = new Configuration();
if(args.length > 0) {
String dataStoreClass = args[0];
inStore = DataStoreFactory.
getDataStore(dataStoreClass, Long.class, Pageview.class, conf);
if(args.length > 1) {
dataStoreClass = args[1];
}
outStore = DataStoreFactory.
getDataStore(dataStoreClass,
String.class, MetricDatum.class, conf);
} else {
inStore = DataStoreFactory.getDataStore(Long.class, Pageview.class, conf);
outStore = DataStoreFactory.getDataStore(String.class, MetricDatum.class, conf);
}
Job job = createJob(inStore, outStore, 3);
boolean success = job.waitForCompletion(true);
inStore.close();
outStore.close();
log.info("Log completed with " + (success ? "success" : "failure"));
return success ? 0 : 1;
}
我们可以看到,生成两个输入与输出的dataStore。
下来看一下自定义的LogAnalyticsMapper代码,它扩展自GoraMapper,它的输入为<Long,Pageview>,部分代码如下:
private TextLong tuple;
protected void map(Long key, Pageview pageview, Context context)
throws IOException ,InterruptedException {
Utf8 url = pageview.getUrl();
long day = getDay(pageview.getTimestamp());
tuple.getKey().set(url.toString());
tuple.getValue().set(day);
context.write(tuple, one);
};
而LogAnalyticsReducer也扩展自GoraReducer,部分代码如下:
protected void reduce(TextLong tuple
, Iterable<LongWritable> values, Context context)
throws IOException ,InterruptedException {
long sum = 0L; //sum up the values
for(LongWritable value: values) {
sum+= value.get();
}
String dimension = tuple.getKey().toString();
long timestamp = tuple.getValue().get();
metricDatum.setMetricDimension(new Utf8(dimension));
metricDatum.setTimestamp(timestamp);
String key = metricDatum.getMetricDimension().toString();
metricDatum.setMetric(sum);
context.write(key, metricDatum);
};
我们可以看到,它输出的Value类型是metricDatum,这个就是我们之前在json中定义的数据模型,它会被序列化到数据库中去。
4. 参考
http://gora.apache.org/docs/current/tutorial.html