Question

我首先将这种关系分为赢得胜利者和失败者。

我很难将所有候选人一起加入，并使用两个候选人的姓氏（选举和失败）以及他们的投票总数（只有差异小于10的元组）之间的差异制作元组。

 --load the data 
raw = LOAD '.../data2.csv' USING PigStorage(',') AS  (
        date, type:chararray, parl:int, prov:chararray, riding:chararray, 
        lastname:chararray, firstname:chararray, gender:chararray,
        occupation:chararray, party:chararray, votes:int,
        percent:double, elected:int);

fltrd = FILTER raw by votes > 100 ;
spltrd = SPLIT fltrd INTO won IF elected > 0, lost IF elected == 0;
jnd = JOIN won BY lastname AS lastname_won, lost BY lastname AS lastname_lost;

为了显示差异btw投票这是我的想法，但它不起作用：

jnd = JOIN won BY lastname AS lastname_won, vote AS vote_won, lost BY lastname AS lastname_lost, vote AS vote_lost;

gen = foreach jnd generate lastname_won, lastname_lost,(vote_won - vote_lost) as diffVotes;

Answer 1

究竟是什么不起作用？

乍一看，由于您使用的AS，您的加入可能无法正常工作。你不能重命名连接中的列，你需要一个foreach。使用::来区分两个关系中的字段：

jnd2 = FOREACH jnd GENERATE won::lastname AS lastname_won, won::vote AS vote_won etc.

如果您加入多个列，请更改语法，如下所示：

jnd = JOIN won BY (lastname, vote), lost BY (lastname, vote);

加入两个关系（候选人）并显示投票之间的差异？

1 个答案: