每个小组中的hadoop pig百分比

Question

我有一个文件如下

姓名得分
约翰aa 约翰aa 约翰aa 约翰bb

mary cc
mary cc
玛丽dd

我想按每个人的分数输出他们的分数百分比 所以它看起来像这样

约翰aa 75
约翰bb 25
玛丽cc 66.6
玛丽dd 33.3

John有3 aa和1 bb，所以aa％= 75，bb％= 25 我想在Hadoop猪中做，请帮助，谢谢

-Troy

Answer 1

你能试试吗？

输入：file.dat

john aa
john aa
john aa
john bb
mary cc
mary cc
mary dd

代码：

A = LOAD 'file.dat' USING PigStorage(' ') as (name:chararray,score:chararray);
N = CUBE A BY CUBE(name,score);      
N2 = FOREACH N GENERATE FLATTEN(group) AS (name,score), ((float)COUNT_STAR(cube)) As (totcnt:float); 
N3 = FILTER N2 BY name!='null';
N4 = GROUP N3 BY name; 
N5 = FOREACH N4 {                                                                        
 fil = order N3 BY score;                                                                 
 fil1 = LIMIT fil 1;                                                                      
 fil2 = FILTER N3 BY score!='null';                                                       
 generate FLATTEN(fil2) AS (name:chararray,score:chararray,indcount:float),FLATTEN(fil1.totcnt) as (totcnt:float);
 }
N6 = FOREACH N5 GENERATE name,score,(indcount/totcnt)*100;
DUMP N6;

输出：

(john,aa,75.0)
(john,bb,25.0)
(mary,cc,66.66667)
(mary,dd,33.333336)