apache-pig - 使用 group by 和 MAX 函数的 Pig Latin 查询

Question

给定表格：

Place(name, province, population, mayorid)

您将如何用 Pig Latin 编写以下查询？为每个省返回人口最多的地方。您的结果集应该包含省名、地名和该地的人口。

Answer 1

没有测试过这个，但类似

places = LOAD 'placesInput' AS (name, province, population, mayorid);
placesProjected = FOREACH places GENERATE name,province,population;
placesGrouped = GROUP placesProjected by province;
biggestPlaces = FOREACH placesGrouped {
    sorted = ORDER placesProjected by population DESC;
    maxPopulation = LIMIT sorted 1;
    GENERATE group as province, FLATTEN(maxPopulation.name) as name, FLATTEN(maxPopulation.population) as population;
};

应该工作。