hadoop - 想要在使用 pig 的记录中找到最大记录

Question

我想找到在 pig 对阵每支球队中得分最多的球员。

Input : Inputs are in the below fashion

Sachin 100 KXIP Hyderabad 1991 sehwag 150 KXIP Hyderabad 1991 Sehwag 100 MI Mumbai 2011 Kohli 0 CSK Chennai 2014 Dhoni 150 MI Hyderabad 1991 Sachin 32 PW Chennai 2014 Dhoni 150 MI Mumbai 2011 我的实现: record1= LOAD 'ipl.txt' using PigStorage(' ') as (name:chararray,runs:int,team:chararray,loc:chararray,year:int); record2 = GROUP record1 by team as team; record3 = FOREACH record2 GENERATE group,MAX(record1.runs) as mx; record4= ORDER record3 by mx ASC; DUMP record4;

输出:
(PW，32)
(KXIP，150)
(MI 150)

但是以以下方式期待结果。
Sachin PW 32钦奈2014

Answer 1

record1= LOAD 'ipl.txt' using PigStorage(' ') as    (name:chararray,runs:int,team:chararray,loc:chararray,year:int);
record2 = GROUP record1 by team;
record3 = FOREACH record2 GENERATE group,MAX(record1.runs) as mx;
record4 = JOIN record3 by (mx,group) LEFT OUTER, record1 by (runs,team);
record5 = FOREACH record4 GENERATE record1::name as name, record1::team as team, record3::mx as mx, record1::year as year;
record6= ORDER record5 by mx ASC;
DUMP record6;

产生以下结果

(Kohli,CSK,0,2014)
(Sachin,PW,32,2014)
(sehwag,KXIP,150,1991)
(Dhoni,MI,150,1991)
(Dhoni,MI,150,2011)

请注意，Dhoni有两条记录，这是因为他两次获得150分。如果要删除，则需要根据需要选择最早或最近的年份。