如何找到具有红移条件的不同行中相同列值之间的差异？

Question

我有一个聊天室数据集，其中包含用户加入和离开房间的活动。一个用户可以多次加入和离开房间。我需要找出每个用户在每个房间花费的时间。

  room_id user_id event   time  
  1       1       join    2021-12-10 09:00:00 
  1       2       join    2021-12-10 09:10:00 
  1       3       join    2021-12-10 09:20:00 
  1       2       leave   2021-12-10 09:30:00 
  1       1       leave   2021-12-10 09:40:00 
  1       1       join    2021-12-10 09:50:00 
  1       3       leave   2021-12-10 10:00:00 
  1       1       leave   2021-12-10 10:10:00

我需要的是：

  room_id user_id time_spend
  1       1       60
  1       2       20
  1       1       30

我是红移新手。我尝试使用滞后功能，但无法添加加入和离开等条件。有人可以帮我解决这个问题吗？

Answer 1

下面是基于您的样本数据的开始：

create table test (room_id int, user_id int, event varchar(16), event_time  timestamp);

insert into test values  
(1,      1,      'join', '2021-12-10 09:00:00'),
(1,      2,      'join', '2021-12-10 09:10:00'), 
(1,      3,      'join', '2021-12-10 09:20:00'), 
(1,      2,      'leave', '2021-12-10 09:30:00'), 
(1,      1,      'leave', '2021-12-10 09:40:00'), 
(1,      1,      'join', '2021-12-10 09:50:00'), 
(1,      3,      'leave', '2021-12-10 10:00:00'), 
(1,      1,      'leave', '2021-12-10 10:10:00')
;

select * from test order by user_id, event_time;

select room_id, user_id, sum(time_spent) as time_spent
from (
    select room_id, user_id, event_time, decode(event,'join',event_time) as join_time, 
                lead(decode(event,'leave',event_time)) ignore NULLS OVER (partition by room_id, user_id order by event_time) as leave_time,
                datediff(min, join_time, leave_time) as time_spent
    from test)
group by room_id, user_id
order by room_id, user_id;

如何找到具有红移条件的不同行中相同列值之间的差异？

How to find the difference between same column values in different rows with condition in redshift?

sql

amazon-redshift