lambda表达式的复合排序，内存分页，分组后求和，异步查询后合并。

lambda表达式现在属于程序员必备的编码技巧了。

简单的map，filter，分组等就不说了。

lambda表达式复合排序：
使用thenComparing

复合排序

public static void main(String[] args) throws Exception {
        List<SysUser> list = new ArrayList<>();
        list.add(new SysUser(1L,"张三",2L,"admin",new Date()));
        list.add(new SysUser(2L,"李莫愁",3L,"common",new Date()));
        list.add(new SysUser(3L,"梅超风",2L,"admin",new Date()));
        list.add(new SysUser(4L,"张三丰",4L,"user",new Date()));
        list.add(new SysUser(5L,"张无忌",3L,"common",new Date()));

        /**
         * 按照roleId 降序排列， 如果roleId相同，则按照UserId升序排列.
         */
        List<SysUser> collect = list.stream()
                .sorted(
                        Comparator.comparing(SysUser::getRoleId)
                                .reversed()
                                .thenComparing(Comparator.comparing(SysUser::getUserId)))
                .collect(Collectors.toList());
        
    }

这样排序出来类似于sql中的 ORDER BY roleId DESC,userId ASC 这种了。

内存分页

有时候分页在sql层面做不了，比如有些字段是统计出来的，统计后又根据其他条件进行查询。

public static void main(String[] args) throws Exception {
        List<SysUser> list = new ArrayList<>();
        list.add(new SysUser(1L,"张三",2L,"admin",new Date()));
        list.add(new SysUser(2L,"李莫愁",3L,"common",new Date()));
        list.add(new SysUser(3L,"梅超风",2L,"admin",new Date()));
        list.add(new SysUser(4L,"张三丰",4L,"user",new Date()));
        list.add(new SysUser(5L,"张无忌",3L,"common",new Date()));

        int pageNum = 1;    //当前页
        int pageSize = 3;   //一页显示3条

        Page<SysUser> page = new Page<>(pageNum, pageSize);
        page.setTotal(list.size());
        
        List<SysUser> collect = list.stream()
                .skip((pageNum - 1) * pageSize) //使用skip 跳过指定条数
                .limit(pageSize)                //使用limit取指定的条数
                .collect(Collectors.toList());  //搜集起来
        
        page.setRecords(collect);
        
    }

求和

public static void main(String[] args) throws Exception {
        List<SysUser> list = new ArrayList<>();
        list.add(new SysUser(1L,"张三",2L,"admin",new Date()));
        list.add(new SysUser(2L,"李莫愁",3L,"common",new Date()));
        list.add(new SysUser(3L,"梅超风",2L,"admin",new Date()));
        list.add(new SysUser(4L,"张三丰",4L,"user",new Date()));
        list.add(new SysUser(5L,"张无忌",3L,"common",new Date()));
		
        Long idSum = list.stream().map(SysUser::getUserId).reduce(0L,Long::sum);

    }

异步查询后合并

异步查询使用到CompletableFuture和Stream.of

public static void main(String[] args) throws Exception {
        List<SysUser> list = new ArrayList<>();
        list.add(new SysUser(1L,"张三",2L,"admin",new Date()));
        list.add(new SysUser(2L,"李莫愁",3L,"common",new Date()));
        list.add(new SysUser(3L,"梅超风",2L,"admin",new Date()));
        list.add(new SysUser(4L,"张三丰",4L,"user",new Date()));
        list.add(new SysUser(5L,"张无忌",3L,"common",new Date()));

        //在list中异步过滤出 角色id大于2的数据
        CompletableFuture<List<SysUser>> future = CompletableFuture.supplyAsync(() -> {
            return list.stream().filter(sysUser -> sysUser.getRoleId() > 2).collect(Collectors.toList());
        });
        //异步future 完成后 继续过滤出 名称中包含张三的数据
        CompletableFuture<List<SysUser>> async = future.thenApplyAsync(list1 -> {
            return list1.stream().filter(sysUser -> sysUser.getUserName().contains("张三")).collect(Collectors.toList());
        });

        //在list中异步过滤出用户类型为admin的数据
        CompletableFuture<List<SysUser>> completableFuture = CompletableFuture.supplyAsync(() -> {
            return list.stream().filter(sysUser -> sysUser.getUserType().equals("admin")).collect(Collectors.toList());
        });

        //主线程等待 异步处理合并结果
        List<Object> futureList = Stream.of(async, completableFuture).map(CompletableFuture::join).collect(Collectors.toList());

        List<SysUser> list1 = (List<SysUser>)futureList.get(0);
        List<SysUser> list2 = (List<SysUser>)futureList.get(1);
        /**
         * list1为角色id大于2，且名称里包含张三的
         * list2为用户类型为admin的数据。
         * 异步获取可以提高接口响应时间
         */
        
        //其他处理...

    }