ForkJoinPool process Arraylist and return two arrays

i have a huge List<String[]> like about 500k elements validation of it takes too long - 35-40 sec validation looks like this

   Iterator<String[]>iterator=parser.iterate(request.getInputStream()).iterator();
    List<String[]> list =new ArrayList<>();
    List<NotValidRow>badList=new ArrayList<>();
    while (iterator.hasNext()){
      var tmp=iterator.next();
      if(tmp.length!=2)continue;
      if (tmp[0] == null || !SKIP_PATTERN.matcher(tmp[0]).matches()) {
        badList.add(new NotValidRow(tmp[0], tmp[1], NotValidRowReason.NOT_VALID_EMAIL));
      }
      if(tmp[1]==null || tmp[1].isBlank()){
        badList.add(new NotValidRow(tmp[0],tmp[1],NotValidRowReason.EMPTY_NAME));
      }
      list.add(tmp);
    }

i think its possible to do it faster with fork join pool but i dnt know how to do it, could you guys help me wtih that

You can use Stream parallel processing, however, you'll have to sneak out the bad list in a thread-safe manner: for example:

var spliterator = Spliterators.spliteratorUnknownSize(iterator, 0);

var badQueue = new ConcurrentLinkedQueue<NotValidRow>();

List<String[]> list = StreamSupport.stream(spliterator, true)
    .filter(tmp -> {
        if (tmp.length != 2) {
            return false;
        }
        if (tmp[0] == null || !SKIP_PATTERN.matcher(tmp[0]).matches()) {
            badQueue.offer(new NotValidRow(tmp[0], tmp[1], NotValidRowReason.NOT_VALID_EMAIL));
            return false;
        }
        if (tmp[1] == null || tmp[1].isBlank()){
            badQueue.offer(new NotValidRow(tmp[0], tmp[1], NotValidRowReason.EMPTY_NAME));
            return false;
        }
        return true;
    })
    .toList();

List<NotValidRow> badList = new ArrayList<>(badQueue);

Edit Apparently, the OP didn't mean to include the bad entries in the good list, so I've updated the answer to filter out the bad entries.