a = load '/user/hadoop/three_tuple_1.txt' as (line:int,col:int,value:int);
b
= load '/user/hadoop/three_tuple_2.txt' as (line:int,col:int,value:int);
c = group a by line;
d = group b by col;
e = cross c,d;
f = foreach e generate c::group,d::group,flatten(c::a),flatten(d::b);
g = filter f by $3 == $5;
h = foreach g generate c::group,d::group,(c::a::value*d::b::value);
i = group h by (c::group,d::group);
j = foreach i generate flatten(group),SUM(h.$2);
dump j;

本文详细介绍了一个使用大数据处理框架实现的数据处理流程。通过加载两个不同文件中的三元组数据,进行分组、交叉连接、过滤和聚合操作,最终生成汇总结果。此流程展示了如何有效地处理大规模数据集,并提供了具体的操作步骤。
1466

被折叠的 条评论
为什么被折叠?



