I have a simple live-case scenario on the SQL Server 2014. I have a table with orders and their quanities:
我在SQL Server 2014上有一个简单的实时案例。我有一个包含订单及其数量的表格:
declare @qty_to_distribute int = 50
create table orders (id int, priority int, qty int)
insert into orders (id, priority, qty) values
(1, 1, 10),
(2, 2, 30),
(3, 3, 20),
(4, 4, 5)
I have a value, for example 50, which i have to distribute between orders, based on their priority Iiterating through the orders
我有一个值,例如50,我必须根据他们的优先顺序通过订单在订单之间分配
- if the amount left is greater then quanity of the current order, the order will be realized, and the
@qty_to_distribute
reduced by the qty of the order, - if not - the order is cancelled.
如果剩余金额大于当前订单的数量,则将实现订单,并且@qty_to_distribute减少订单的数量,
如果不是 - 订单被取消。
Looking at the sample data - the orders 1, 2 and 4 will be realized and #3 cancelled.
查看样本数据 - 将实现订单1,2和4,并取消#3。
Is there a simple, effective and non-cursor solution, preferably based on window functions?
是否有一个简单,有效和非游标的解决方案,最好是基于窗口函数?
This looks promising, but can't figure out where to put the conditions:
这看起来很有希望,但无法弄清楚在哪里放置条件:
select id,
sum(qty) OVER (ORDER BY priority asc ROWS UNBOUNDED PRECEDING) as qtysum
from orders
id qtysum
1 10
2 40
3 60 *
4 65 *
http://sqlfiddle.com/#!6/05fa2/2/0
2 个解决方案
#1
6
Try this:
create table o (id int, priority int, qty int)
insert into o (id, priority, qty) values
(1, 1, 10),
(2, 2, 30),
(3, 3, 20),
(4, 4, 5),
(5, 5, 1),
(6, 6, 10)
with cte1 as(select *, row_number() over(order by priority) as rn from o),
cte2 as(
select top 1 id, priority, qty, rn,
case when qty > 50 then 1 else 0 end as ToBeCanceled
from cte1 order by rn
Union all
select cte1.id, cte1.priority, case when cte2.qty + cte1.qty > 50 then cte2.qty
else cte2.qty + cte1.qty end as qty,
cte1.rn,
case when cte2.qty + cte1.qty > 50 then 1 else 0 end as ToBeCanceled
from cte1
join cte2 on cte1.rn = cte2.rn + 1)
select id, priority, ToBeCanceled from cte2
option(maxrecursion 0)
Output:
id ToBeCanceled
1 0
2 0
3 1
4 0
5 0
6 1
#2
4
UPDATED thanks to Giorgi Nakeuri
更新感谢Giorgi Nakeuri
WITH Intermediate AS (
SELECT
Id,
ROW_NUMBER() OVER (ORDER BY Priority) as Priority,
qty
FROM Orders),
Result AS
(
SELECT
Id,
Priority,
CASE when qty < 50 THEN 0 ELSE qty END as Residual,
CASE when qty < 50 THEN 50 - qty ELSE 50 END as Remaining
FROM Intermediate
WHERE Priority = 1
UNION ALL
SELECT
ORD.Id,
ORD.Priority,
CASE when ORD.qty < Result.Remaining THEN 0 ELSE qty END as Residual,
CASE when ORD.qty < Result.Remaining THEN Result.Remaining - ORD.qty ELSE Result.Remaining END as Remaining
FROM Intermediate ORD INNER JOIN Result ON ORD.Priority = Result.Priority + 1
)
select Id, Priority, Residual, Remaining from result
#1
6
Try this:
create table o (id int, priority int, qty int)
insert into o (id, priority, qty) values
(1, 1, 10),
(2, 2, 30),
(3, 3, 20),
(4, 4, 5),
(5, 5, 1),
(6, 6, 10)
with cte1 as(select *, row_number() over(order by priority) as rn from o),
cte2 as(
select top 1 id, priority, qty, rn,
case when qty > 50 then 1 else 0 end as ToBeCanceled
from cte1 order by rn
Union all
select cte1.id, cte1.priority, case when cte2.qty + cte1.qty > 50 then cte2.qty
else cte2.qty + cte1.qty end as qty,
cte1.rn,
case when cte2.qty + cte1.qty > 50 then 1 else 0 end as ToBeCanceled
from cte1
join cte2 on cte1.rn = cte2.rn + 1)
select id, priority, ToBeCanceled from cte2
option(maxrecursion 0)
Output:
id ToBeCanceled
1 0
2 0
3 1
4 0
5 0
6 1
#2
4
UPDATED thanks to Giorgi Nakeuri
更新感谢Giorgi Nakeuri
WITH Intermediate AS (
SELECT
Id,
ROW_NUMBER() OVER (ORDER BY Priority) as Priority,
qty
FROM Orders),
Result AS
(
SELECT
Id,
Priority,
CASE when qty < 50 THEN 0 ELSE qty END as Residual,
CASE when qty < 50 THEN 50 - qty ELSE 50 END as Remaining
FROM Intermediate
WHERE Priority = 1
UNION ALL
SELECT
ORD.Id,
ORD.Priority,
CASE when ORD.qty < Result.Remaining THEN 0 ELSE qty END as Residual,
CASE when ORD.qty < Result.Remaining THEN Result.Remaining - ORD.qty ELSE Result.Remaining END as Remaining
FROM Intermediate ORD INNER JOIN Result ON ORD.Priority = Result.Priority + 1
)
select Id, Priority, Residual, Remaining from result