pigscript error not calculation max

Question

我遇到了一个猪脚本的问题，我尝试了很多不同的方法。任何人都可以指出我究竟做错了什么？它应该是非常简单的，我试图在计算平均值后得到最大值。

a = LOAD 'default.books' using org.apache.hcatalog.pig.HCatLoader(); 
b = LOAD 'default.book_rating' using org.apache.hcatalog.pig.HCatLoader(); 

books_and_ratings = join a by isbn, b by isbn; 

by_isbn = GROUP books_and_ratings BY (a::isbn); 

DESCRIBE by_isbn; 

average_book_rating = FOREACH by_isbn 
     GENERATE books_and_ratings.book_title, books_and_ratings.a::isbn as isbn1, 
     books_and_ratings.book_author, books_and_ratings.publisher, 
     AVG(books_and_ratings.book_rating) as AVG_RATING; 

DESCRIBE average_book_rating; 

group_avg = GROUP average_book_rating ALL; 

DESCRIBE group_avg; 

max_avg_rating = FOREACH group_avg 
    GENERATE FLATTEN average_book_rating.a::book_title, isbn1, 
      average_book_rating.a::book_author, average_book_rating.a::publisher, MAX(AVG_RATING); 

dump max_avg_rating; 

解析失败：不匹配的输入 'average_book_rating' 期待LEFT_PAREN

Answer 1

你可以尝试这样的。

max_avg_rating = ORDER average_book_rating BY AVG_RATING DESC; 
top_most_rating = LIMIT max_avg_rating 1; 
dump top_most_rating; 

Answer 2

看到阎王最新评论后（“可以有多种书籍最高平均等级”），我想你需要另一组，第一个，获得通过书号哪些群体的收视率，你想要的东西之后。

开始是这样的：由AVG_RATING

grouped_rating = GROUP average_book_rating;

然后你可以使用像@ Sivasakthi代码：

ordered_avg_rating = ORDER BY grouped_rating DESC组;
top_most_rating = LIMIT ordered_avg_rating 1;
dump top_most_rating;

这样一来，如果有与平等，最高收视多个结果，top_most_rating将所有的信息接受这个最高等级的书袋。当然，如果你不想把它作为一个包，你可以把它设计得更方便些。

UPDATE：

这是我怎么会改变上面的代码。有一件事情不是纯粹的功能，我会首先将评分平均，然后加入书籍/作者信息 - 这会更好地表现明智，否则你会增加评分的大小（其中有很多）时，他们去了。

所以它看起来像这样：

-- assume a: book_title, isbn, book_author, publisher (and maybe more, which we'll ignore) 
    a = LOAD 'default.books' using org.apache.hcatalog.pig.HCatLoader(); 

    -- assume b: isbn, book_rating (and maybe more, which we'll ignore) 
    b = LOAD 'default.book_rating' using org.apache.hcatalog.pig.HCatLoader(); 

    by_isbn = GROUP b BY isbn; 

    average_book_rating = FOREACH by_isbn GENERATE AVG(b.book_rating) AS AVG_RATING, group AS isbn; 

    group_avg = GROUP average_book_rating BY AVG_RATING; 

    ordered_avg_rating = ORDER group_avg BY group DESC; 

    top_most_rating = LIMIT ordered_avg_rating 1; 

    b = FOREACH top_most_rating GENERATE flatten(average_book_rating); 

    -- now add the book information 

    books_and_ratings = JOIN a BY isbn, b BY isbn; 

    books_and_ratings = FOREACH books_and_ratings GENERATE a::book_title AS title, a::isbn AS isbn, a::book_author AS author,a::publisher AS publisher, b::average_book_rating::AVG_RATING AS max_rating; 

希望这个作品送给你。