我想了解如何使用MapReduce模型聚合多个字段。
例如,如果我有一个这样的数据文件:
id, site, name, qty, price
00, testA, NameA,1,1
01, testB,NameA,2,3
02, testB,NameB,5,7
并希望在MapReduce上实现此聚合:
select site,name, (qty*price) as total
from PO where name ='NameA'
group by site,name,total
order by site;
我将如何去做呢。
我能够按网站(密钥),总计(值)进行汇总,但不知道如何包含名称列。
我需要了解如何使用MapReduce中的多个字段。有一个我可以看的例子吗?或者我需要使用Hbase吗?
答案 0 :(得分:1)
您可以实施WritableComparable
并使用多个字段创建自己的CompositeKey
,例如:
public static class CompositeKey implements WritableComparable<CompositeKey> {
public final Text site;
public final Text name;
public final LongWritable total;
public CompositeKey(Text site, Text name, LongWritable total) {
this.site = site;
this.name = name;
this.total = total;
}
@Override
public int compareTo(CompositeKey o) {
int siteCmp = site.compareTo(o.site);
if (siteCmp != 0) {
return siteCmp;
}
int nameCmp = name.compareTo(o.name);
if (nameCmp != 0) {
return nameCmp;
}
return total.compareTo(o.total);
}
@Override
public boolean equals(Object o) {
if (this == o) return true;
if (o == null || getClass() != o.getClass()) return false;
CompositeKey that = (CompositeKey) o;
if (name != null ? !name.equals(that.name) : that.name != null) return false;
if (site != null ? !site.equals(that.site) : that.site != null) return false;
if (total != null ? !total.equals(that.total) : that.total != null) return false;
return true;
}
@Override
public int hashCode() {
int result = site != null ? site.hashCode() : 0;
result = 31 * result + (name != null ? name.hashCode() : 0);
result = 31 * result + (total != null ? total.hashCode() : 0);
return result;
}
@Override
public void write(DataOutput dataOutput) throws IOException {
site.write(dataOutput);
name.write(dataOutput);
total.write(dataOutput);
}
@Override
public void readFields(DataInput dataInput) throws IOException {
site.readFields(dataInput);
name.readFields(dataInput);
total.readFields(dataInput);
}
}