使用 Java Streams 返回单词出现的句子的计数和列表

Question

I'm stuck trying to get in what sentences each word appears. The entry would be a list of sentences

Question, what kind of wine is best? 
White wine.
A question

and the output would be

// format would be: word:{count: sentence1, sentence2,...}
a:{1:3} 
wine:{2:1,2} 
best:{1:1} 
is:{1:1} 
kind:{1:1} 
of:{1:1} 
question:{2:1,3} 
what:{1:1}
white:{1:2}

This is what I get so far:

static void getFrequency(List<String> inputLines) {
  List<String> list = inputLines.stream()
     .map(w -> w.split("[^a-zA-Z0-9]+"))
     .flatMap(Arrays::stream)
     .map(String::toLowerCase)
     .collect(Collectors.toList());

   Map<String, Integer> wordCounter = list.stream()
     .collect(Collectors.toMap(w -> w, w -> 1, Integer::sum));
}

With that I'm only getting the count of times each word appears in all the sentences, but I need to get also the list of sentences where the word appears. It looks like maybe to get the id of sentences I can use IntStream.range , something like this:

 IntStream.range(1, inputLines.size())
          .mapToObj(i -> inputLines.get(i));

But I'm not sure if that is the best way to do it, I'm new with Java

Answer 1

You can use a grouping collector to compute a word to index list map. Here's an example:

private static Map<String, List<Integer>> getFrequency(List<String> inputLines) {
    return IntStream.range(0, inputLines.size())
            .mapToObj(line -> Arrays.stream(inputLines.get(line)
                 .split("[^a-zA-Z0-9]+"))
                 .map(word -> new SimpleEntry<>(word.toLowerCase(), line + 1)))
            .flatMap(Function.identity())
            .collect(Collectors.groupingBy(Entry::getKey, 
                  Collectors.mapping(Entry::getValue, Collectors.toList())));
}

With your test data, I get

{a=[3], what=[1], white=[2], question=[1, 3], kind=[1], 
 of=[1], best=[1], is=[1], wine=[1, 2]}

The count is easy to infer from the list size, so there should be no need for an additional class.