Filter 依靠多个列上的GROUPBY并获取原始数据集

剧本

2, cornflakes, Regular,General Mills, 12    
3, cornflakes, Mixed Nuts, Post, 14  
4, chocolate syrup, Regular, Hersheys, 5   
5, chocolate syrup, No High Fructose, Hersheys, 8  
6, chocolate syrup, Regular, Ghirardeli, 6  
7, chocolate syrup, Strawberry Flavor, Ghirardeli, 7

---

filter_数据将为您提供

巧克力糖浆，常规的

。将filter_数据与原始数据集中的项连接起来，键入并获得所需的结果

4, chocolate syrup, Regular, Hersheys, 5
6, chocolate syrup, Regular, Ghirardeli, 6

data\u grp=按（项目、类型）分组数据；
data_cnt=FOREACH data_grp生成展平（组）为（项目、类型），计数（数据）为总计；
filter_data=按总和<2过滤数据；
o_data=按（项目、类型）联接数据，按（$0，$1）过滤数据；
最终数据=FOREACH o_数据生成$0..$4；
转储最终数据；

man，把你的问题写清楚……无法生成合理的计划。嵌套异常：org.apache.pig.backend.executionengine.ExecutionException:错误1070:无法使用导入解析计数：[，java.lang.，org.apache.pig.builtin.，org.apache.pig.impl.builtin.]计数应为大写

4, chocolate syrup, Regular, Hersheys, 5
6, chocolate syrup, Regular, Ghirardeli, 6

data_grp = GROUP data BY (item, type);
data_cnt = FOREACH data_grp GENERATE FLATTEN (group) AS(item, type), COUNT(data) as total; 
filter_data = FILTER data_cnt BY total < 2;
o_data = JOIN data BY (item,type),filter_data BY ($0,$1);
final_data = FOREACH o_data GENERATE $0..$4;
DUMP final_data;