How to groupBy and Collect with Java 8?

Question

I have a list of elements, let's call it "keywords", like this:

public class Keyword {
    Long id;
    String name;
    String owner;
    Date createdTime;
    Double price;
    Date metricDay;
    Long position;
}

The thing is that there is a keyword for every single day. For example:

Keyword{id=1, name="kw1", owner="Josh", createdTime="12/12/1992", price="0.1", metricDay="11/11/1999", position=109}
Keyword{id=1, name="kw1", owner="Josh", createdTime="12/12/1992", price="0.3", metricDay="12/11/1999", position=108}
Keyword{id=1, name="kw1", owner="Josh", createdTime="12/12/1992", price="0.2", metricDay="13/11/1999", position=99}
Keyword{id=2, name="kw2", owner="Josh", createdTime="13/12/1992", price="0.6", metricDay="13/11/1999", position=5}
Keyword{id=2, name="kw2", owner="Josh", createdTime="13/12/1992", price="0.1", metricDay="14/11/1999", position=4}
Keyword{id=3, name="kw3", owner="Josh", createdTime="13/12/1992", price="0.1", metricDay="13/11/1999", position=8}

Then, from this list I would like to create a new list with all the metrics from all those different days on one single list. First, I created a class like this:

public class KeywordMetric {
    Double price;
    Date metricDay;
    Long position;
} 

And what I would like to archive is go from the first list, to a structure like this:

public class KeywordMeged {
    Long id;
    String name;
    String owner;
    List<KeywordMetric> metricList;
}

Example of what I expect:

KeywordMerged{id=1, name="kw1", owner="Josh", createdTime="12/12/1992", metricList=[KeywordMetric{price=0.1,metricDay="11/11/1999",position=109},KeywordMetric{price=0.3,metricDay="12/11/1999",position=108},KeywordMetric{price=0.2,metricDay="13/11/1999",position=99}]
KeywordMerged{id=2, name="kw2", owner="Josh", createdTime="13/12/1992", metricList=[KeywordMetric{price=0.6,metricDay="13/11/1999",position=5},KeywordMetric{price=0.1,metricDay="14/11/1999",position=4}]
KeywordMerged{id=3, name="kw3", owner="Josh", createdTime="13/12/1992", metricList=[KeywordMetric{price=0.1,metricDay="13/11/1999",position=8}]

I know how to do this with a lot of loops and mutable varibles, but I can't figure out how to do this with streams and lambda operations. I was able to group all related keywords by Id with this:

Map<Long, List<Keyword>> kwL = kwList.stream()
                    .collect(groupingBy(Keyword::getId))

And I know that with .forEach() I could iterate over that Map, but can't figure out how to make the collect() method of streams pass from List to KeywordMerged.

Answer 1

You can try to use the Collectors.toMap(...) instead. Where:

Keyword::getId is a key mapper function.

KeywordMerged.from(...) performs a transformation: Keyword => KeywordMerged

(left, right) -> { .. } combines metrics for entities with identical ids.

Collection<KeywordMerged> result = keywords.stream()
    .collect(Collectors.toMap(
        Keyword::getId,
        k -> KeywordMerged.from(k), // you can replace this lambda with a method reference
        (left, right) -> {
            left.getMetricList().addAll(right.getMetricList());
            return left;
        }))
    .values();

A transformation method might look something like this:

public class KeywordMerged {
    public static KeywordMerged from(Keyword k) {            
        KeywordMetric metric = new KeywordMetric();
        metric.setPrice(k.getPrice());
        metric.setMetricDay(k.getMetricDay());
        metric.setPosition(k.getPosition());

        KeywordMerged merged = new KeywordMerged();
        merged.setId(k.getId());
        merged.setName(k.getName());
        merged.setOwner(k.getOwner());
        merged.setMetricList(new ArrayList<>(Arrays.asList(metric)));
        return merged;
    }
}

I think you've got the basic idea. So, refactor according to your needs...

Answer 2

A slightly different approach. First you collect the Map of keywords grouped by id:

Map<Integer, List<Keyword>> groupedData = keywords.stream()
    .collect(Collectors.groupingBy(k -> k.getId()));

Further you convert your map to the list of desired format:

List<KeywordMerged> finalData = groupedData.entrySet().stream()
    .map(k -> new KeywordMerged(k.getValue().get(0).getId(),
        k.getValue().stream()
            .map(v -> new KeywordMetric(v.getMetricDay(), v.getPrice(), getPosition()))
            .collect(Collectors.toList())))
    .collect(Collectors.toList());

This will work on the grouped data, but transforming the map it will create KeywordMerged object, which as argument will receive id (you can extent it further yourself) and converted to List<KeywordMetric> previously grouped by ID data.

EDIT: I believe with some extraction to methods you can make it look much nicer :)