我正在尝试做类似于this的事情,但我需要通过公共字段进行分组的额外复杂性
我有两个表,一个用于competition_users
(对于那些在比赛中竞争)和另一个competition_times
(用于存储他们最快的时间),但在同一个表中有许多不同的比赛。
我在rank
表中创建了一个新列competition_times
,现在需要为现有用户运行更新以显示他们在竞争中的排名,我正在尝试修改在其他SO回答没有成功
表格如下......
competition_users:
competition_user_id
,competition_id
competition_times:
competition_time_id
,competition_user_id
,time_in_seconds
,rank
我不确定是否可以使用GROUP BY
?或者如果有另一种方式,到目前为止我正在尝试这样的事情......
UPDATE competition_times
SET rank = r.rnk
FROM (
SELECT competition_time_id
, dense_rank() OVER (ORDER BY time_in_seconds ASC) AS rnk
FROM competition_times, competition_users
WHERE competition_times.competition_user_id = competition_users.competition_user_id
GROUP BY competition_users.competition_id
) r
WHERE competition_times.competition_time_id = r.competition_time_id
使用postgreSQL 9
由于
答案 0 :(得分:5)
你需要在窗口函数中使用PARTITION BY
子句来保持比赛之间的排名。
UPDATE competition_times
SET
rank = r.rnk
FROM (
SELECT competition_time_id, dense_rank() OVER (PARTITION BY competition_id ORDER BY time_in_seconds ASC) AS rnk
FROM competition_times
INNER JOIN competition_users
ON competition_times.competition_user_id = competition_users.competition_user_id
) r
WHERE competition_times.competition_time_id = r.competition_time_id
以下示例数据:
create table competition_users
(competition_user_id int, competition_id int);
create table competition_times
(competition_time_id int, competition_user_id int, time_in_seconds int, rank int);
insert into competition_users values
(1,1),(2,1),(3,1),(4,1),(5,2),(6,2);
insert into competition_times values
(1,1,10,null),
(2,2,20,null),
(3,3,15,null),
(4,4,15,null),
(5,5,10,null),
(6,6,7,null);
我在竞争时间得到了结果:
competition_time_id competition_user_id time_in_seconds rank
1 1 10 1
2 2 20 3
3 3 15 2
4 4 15 2
5 5 10 2
6 6 7 1
第1-4行在一场比赛中,第5行和第6行在另一场比赛中。