先区分，然后按类别分组。假设已将数据加载到关系a中。
选择加载后的两列。
区别关系
按类别分组
统计每个类别的职业

B = FOREACH A GENERATE Occupation as Occupation,Category as Category;
C = DISTINCT B;
D = GROUP C BY $1; 
E = FOREACH D GENERATE group,COUNT(C.Occupation); 
DUMP E;

试试这个：

x= load '<data>' using PigStorage('\t') as (occupation:chararray,category:chararray,name:chararray);

 x_grouped= group x by category;

x_grouped_distinct= foreach x_grouped { cat= distinct $1.occupation; generate $0, cat, COUNT(cat);}; 

dump x_grouped_distinct;