I’ve got 2 Solutions:

#1 Solution (Multiple for each loops)

// entries will be sorted by your comparator NavigableMap<Date, List<GpsDTO>> map = new TreeMap<yourComparator>(); // if there is a list already, it will be used; otherwise, a new one will be created gpsDtoList.forEach(dto ->      map.computeIfAbsent(dto.getEreigniszeit(), new ArrayList<>())                             .add(dto));  // iterate the relevant lists to set the respective markers map.firstEntry().forEach(dto -> dto.setOldest()); if (map.size() > 1) {     map.lastEntry().forEach(dto -> dto.setNeweset()); } 

In my understand this is O(3n)

and this is

#2 Solution (single for loop)

Date youngestDate = null; Date oldestDate = null;  Collections.sort(gpsDtoList, new Comparator<T>() {              public int compare(T o1, T o2) {                  return -o1.getEreigniszeit().compareTo(o2.getEreigniszeit());              }         });  if (gpsDtoList != null && gpsDtoList.size() > 1) {             youngestDate = gpsDtoList.get(0).getEreigniszeit();             oldestDate = gpsDtoList.get(gpsDtoList.size() - 1).getEreigniszeit();         }  and the actual setting logic  if (youngestDate != null && oldestDate != null) {                 if (i == 0 || youngestDate.equals(gpsDtoList.get(i).getEreigniszeit())) {                     properties.put("markerStyle", LATEST);                 } else if (oldestDate.equals(gpsDtoList.get(i).getEreigniszeit())) {                     properties.put("markerStyle", OLDEST);                 }              } 

This one uses a count based for loop. Which should be O(n)

Question

What has the better performance here, I’d go with Solution 2, but is it really more performant?