Spark SQL嵌套查询

Question

我有以下Spark sql查询

SELECT count(*),
       channel
FROM channelusage a
WHERE a.starttime>= windowstarttime
  AND a.endtime <= windowendtime
GROUP BY channel

我必须为10个窗口生成这些计数。目前，我使用while循环来生成windowstarttime和windowendtime。

我想做的是-我想在sql查询本身中生成窗口，类似于类似于此的嵌套sql-

SELECT count(*),
       channel
FROM channelusage a
WHERE (nested query logic)
GROUP BY channel

这样我得到的输出与此类似

windowstarttime | windowentime | channel | count
11:00:01        | 11:00:10     | ABC     |2
11:00:11        | 11:00:20     | ABC     |4
11:00:21        | 11:00:30     | NBC     |10
11:00:31        | 11:00:40     | CNN     |5