我有下表
id_val date_val volume_val
1 26-Jan-18 100
1 25-Jan-18 200
1 24-Jan-18 300
1 23-Jan-18 150
2 26-Jan-18 250
2 25-Jan-18 350
2 24-Jan-18 400
3 23-Jan-18 500
我需要通过id估算累积量。这就是我想要的
id_val date_val volume_val cumulative_volume_val
1 26-Jan-18 100 100
1 25-Jan-18 200 300
1 24-Jan-18 300 600
1 23-Jan-18 150 750
2 26-Jan-18 250 250
2 25-Jan-18 350 600
2 24-Jan-18 400 1000
3 23-Jan-18 500 500
这是我尝试的但它不起作用。它总结了一切,并没有按群体分割。我怎样才能改进它?
set @csum := 0;
select id_val, date_val,
(@csum := @csum + volume_val) as cumulative_volume
from temp
group by id_val, date_val
order by id_val, date_val desc
答案 0 :(得分:1)
以下是我根据类似问题的答案改编的解决方案:Does mysql have the equivalent of Oracle's "analytic functions"?
select id_val, date_val,
(case when @id_val != id_val then @sum := volume_val else @sum := @sum + volume_val end) as cumulative_volume,
(case when @id_val != id_val then @id_val := id_val else @id_val end) as _
FROM (SELECT * FROM temp ORDER BY id_val, date_val desc) z
JOIN (SELECT @sum := 0) s
JOIN (SELECT @idval := -1) i