按流java 8中的多个字段分组

Question

I have a class:

    public class PublicationDTO {

     final String publicationName;
     final String publicationID;
     final String locale;
     final Integer views;
     final Integer shares;
}

A need to get sum of views and shares. For test i created a list:

        PublicationDTO publicationDTO1 = new PublicationDTO("Name1", "name1", "CA", 5, 6);
        PublicationDTO publicationDTO2 = new PublicationDTO("Name2", "name2", "US", 6, 3);
        PublicationDTO publicationDTO3 = new PublicationDTO("Name1", "name1", "CA", 10, 1);
        PublicationDTO publicationDTO4 = new PublicationDTO("Name2", "name2", "CA", 2, 3);

        List<PublicationDTO> publicationDTOS = List.of(publicationDTO1, publicationDTO2, publicationDTO3, publicationDTO4);

I want to group objects in list by publicationName, publicationId and locale and get result list like:

List.of(new PublicationDTO("Name1", "name1", "CA", 15, 7),
        new PublicationDTO("Name2", "name2", "CA", 2, 3),
        new PublicationDTO("Name2", "name2", "US", 6, 3));

I found a solution like:

List<PublicationDTO> collect = publicationDTOS.stream()
            .collect(groupingBy(PublicationDTO::getPublicationID))
            .values().stream()
            .map(dtos -> dtos.stream()
                    .reduce((f1, f2) -> new PublicationDTO(f1.publicationName, f1.publicationID, f1.locale, f1.views + f2.views, f1.shares + f2.shares)))
            .map(Optional::get)
            .collect(toList());

but the result not grouped by locale and I'm not sure if it works by publicationId. Please let me know how to properly use collectors in such case?

Answer 1

You are only grouping by getPublicationID :

.collect(groupingBy(PublicationDTO::getPublicationID))

Since those fields you are interested in for grouping are all strings you could concatenate them and use that as a grouping classifier:

.collect(Collectors.groupingBy(p -> p.getPublicationName() + p.getPublicationID() + p.getLocale()))