Multi-stream: build an inverted index with flatMap and Map.entry

Turning documents-to-tags into tags-to-documents is a flatMap that emits a pair per tag, then a groupingBy that flips it. Map.entry is the cheapest pair type in the JDK and needs no record of its own.

Code

record Doc(String id, List<String> tags) {}

var docs = List.of(new Doc("d1", List.of("java", "streams")),
                   new Doc("d2", List.of("streams", "kafka")),
                   new Doc("d3", List.of("java")));

Map<String, List<String>> byTag = docs.stream()
        .flatMap(d -> d.tags().stream().map(t -> Map.entry(t, d.id())))
        .collect(Collectors.groupingBy(Map.Entry::getKey, TreeMap::new,
                 Collectors.mapping(Map.Entry::getValue, Collectors.toList())));

byTag.forEach((tag, ids) -> System.out.println(tag + " -> " + ids));

Output

java -> [d1, d3]
kafka -> [d2]
streams -> [d1, d2]