Java JSONArray和迭代的并行化

Question

I have a piece of code that I would like to parallelize as they are independent operations -

        List<StockQuote> topGainers = new ArrayList<StockQuote>();
        JSONObject jsonObject = (JSONObject)new JSONParser().parse(
                new InputStreamReader(response.getEntity().getContent(), "UTF-8"));
        JSONArray dataArray = (JSONArray) jsonObject.get("data");

        //Parallelize the for loop
        for(int iter=0;iter<dataArray.size();iter++) {
            JSONObject temp = (JSONObject) dataArray.get(iter);
            System.out.println(temp.get("symbol"));
            //getQuote function has network calls. 
            //Serial implementation makes it take too much time
            topGainers.add(this.getQuote((String)temp.get("symbol")));
        }

        return topGainers;

How can this piece of code be parallelized? Is ArrayList and its Add operation threadsafe?

I tried this -

        int size = dataArray.size();
        ForkJoinPool forkJoinPool = new ForkJoinPool(size);
        forkJoinPool.submit(() ->
        IntStream.range(1, size).parallel().filter( (IntPredicate)i -> { 
        JSONObject temp = (JSONObject) dataArray.get(i);
            System.out.println(temp.get("symbol"));
            try {
                System.out.println(temp.get("symbol"));
                topGainers.add(this.getQuote((String)temp.get("symbol")));
            } catch (Exception e) {
                // TODO Auto-generated catch block
                e.printStackTrace();
            }
        return true;}));

I got the topGainers array as empty

Answer 1

ArrayList is not a Synchronized collection. So, ArrayList#add is not thread-safe. Use CopyOnWriteArrayList instead which is thread safe. Java-8 provided us some excellent feature one of those is parallelStream over the collection. So you can take advantage of that. However, you are always free to create your own thread pool and execute the task if it gives you more benefit.

Below example may help you

    CopyOnWriteArrayList<StockQuote> topGainers = new CopyOnWriteArrayList<>();
    dataArray.parallelStream().forEach(e->{
        JSONObject temp = (JSONObject) dataArray.get(iter);
        System.out.println(temp.get("symbol"));
        //getQuote function has network calls. 
        //Serial implementation makes it take too much time
        topGainers.add(this.getQuote((String)temp.get("symbol")));
    });

You cloud also use Collections#synchronizedList .

Answer from the question 1 from comment: dataArray.parallelStream().forEach will iterate over the list in parallel.

Answer to your question 2 from comment :

   dataArray.parallelStream().forEach(e->{
        JSONObject temp = (JSONObject) dataArray.get(iter);
        System.out.println(temp.get("symbol"));
        //getQuote function has network calls. 
        //Serial implementation makes it take too much time
        if (some condition here) {
           throw new YourException("Your exception messge");
         }
        topGainers.add(this.getQuote((String)temp.get("symbol")));
    });