借用医学的那句话,抛开数据量谈性能都是耍流氓,本次的性能测试分为三档,即万级数据、十万级数据、百万级数据
万级数据量
List<User> userList = new ArrayList<>();
Random rand = new Random();
for (int i = 0; i <10000 ; i++) {
User user = new User();
user.setId(rand.nextInt(1000));
user.setCompanyId(rand.nextInt(1000));
userList.add(user);
}
Long startTime = System.currentTimeMillis();
userList.sort((a, b)-> b.getId().compareTo(a.getId()));
System.out.println("List.sort()耗时:"+(System.currentTimeMillis()-startTime)+"ms");
Long startTime1 = System.currentTimeMillis();
userList.stream().sorted(Comparator.comparing(User::getCompanyId)).collect(Collectors.toList());
System.out.println("stream.sort耗时:"+(System.currentTimeMillis() - startTime1)+"ms");
运行结果:
List.sort() 耗时:116ms
stream.sort耗时:27ms
结论:很明显是stream优势非常明显,相差4~5倍!
十万级数据量
List<User> userList = new ArrayList<>();
Random rand = new Random();
for (int i = 0; i <100000 ; i++) {
User user = new User();
user.setId(rand.nextInt(1000));
user.setCompanyId(rand.nextInt(1000));
userList.add(user);
}
Long startTime = System.currentTimeMillis();
userList.sort((a, b)-> b.getId().compareTo(a.getId()));
System.out.println("List.sort()耗时:"+(System.currentTimeMillis()-startTime)+"ms");
Long startTime1 = System.currentTimeMillis();
userList.stream().sorted(Comparator.comparing(User::getCompanyId)).collect(Collectors.toList());
System.out.println("stream.sort耗时:"+(System.currentTimeMillis() - startTime1)+"ms");
运行结果:
List.sort()耗时:150ms
stream.sort耗时:197ms
List.sort()耗时:142ms
parallelStream.sort耗时:193ms
结论:可以看到在十万级数据量stream已经优势不再,并行流也一样,反而List接口的默认方法却优势很突出
百万级数据量
List<User> userList = new ArrayList<>();
Random rand = new Random();
for (int i = 0; i <1000000 ; i++) {
User user = new User();
user.setId(rand.nextInt(1000));
user.setCompanyId(rand.nextInt(1000));
userList.add(user);
}
Long startTime = System.currentTimeMillis();
userList.sort((a, b)-> b.getId().compareTo(a.getId()));
System.out.println("List.sort()耗时:"+(System.currentTimeMillis()-startTime)+"ms");
Long startTime1 = System.currentTimeMillis();
userList.stream().sorted(Comparator.comparing(User::getCompanyId)).collect(Collectors.toList());
System.out.println("stream.sort耗时:"+(System.currentTimeMillis() - startTime1)+"ms")
运行结果:
List.sort()耗时:523ms
stream.sort耗时:824ms
可以发现随着数据量的增大,stream的排序效率越来越差,由此可以得出结论:
在万级数据及以下,优先选择stream排序;
在十万级及以上,那么则需要选择List提供的sort方法进行排序效率更高!