with recursive cte (id, qty) (
select id, qty from mytable
union all
select id, qty - 1 from cte where qty > 1
)
select id from cte order by id
问题:
如何添加将按正确顺序显示的日期(月(,现在它在递归CTE中向后?
with recursive cte (id, qty, Forecast_date)
(
select id, qty, StartDate from mytable
union all
select id, qty - 1, eomonth(dateadd(month, 1,Forecast_date)) from cte where qty > 1 )
select id,Forecast_date from cte order by id
输出:
9/30/2023 0
8/31/2023 1
7/31/2023 2
6/30/2023 3
5/31/2023 4
4/30/2023 5
3/31/2023 6
2/28/2023 7
1/31/2023 8
12/31/2022 9
11/30/2022 10
10/31/2022 11
9/30/2022 12
8/31/2022 13
7/31/2022 14
6/30/2022 15
5/31/2022 16
...
预计增量从2022年5月31日开始,而不是从2023-09-30开始,这应该是最后一次记录
要点:
Id Forecast_date qty
1 5/31/2022 0
2 6/30/2022 1
减去种子中的月数选择:
with recursive cte (id, qty, Forecast_date)
(
select id, qty, eomonth(dateadd(month, -qty,StartDate))
from mytable
union all
select id, qty - 1, eomonth(dateadd(month, 1,Forecast_date))
from cte
where qty >= 1
)
select *
from cte
order by id, qty
或者添加另一列,从数量0开始,向上计数而不是向下计数:
with recursive cte (id, qty, Forecast_date, stopat)
(
select id, 0, StartDate, qty
from mytable
union all
select id, qty + 1, eomonth(dateadd(month, 1,Forecast_date)), stopat
from cte
where qty < stopat
)
select id, qty, Forecast_date
from cte
order by id, qty
从生成递增序列开始,然后在外部查询中进行计算可能会更简单:
with recursive cte (id, qty, forecast_date, lvl) (
select id, qty, datefromparts(year(forecast_date), month(forecast_date), 1), 0 lvl from mytable
union all
select id, qty, forecast_date, lvl + 1 from cte where lvl < qty
)
select
id,
qty - lvl as dty,
eomonth(dateadd(month, -lvl, forecast_date)) new_forecast_date
from cte
order by id, new_forecast_date
月末计算很棘手:你不能通过添加月份从一个月末跳到另一个月末,因为并非所有月份的天数都相同。因此,查询使用开始月份的开始进行计算,然后使用eomonth()
。