SQL Server：高度な更新クエリ

Question

推定日次在庫量を計算するSQLの更新クエリについてのヘルプが必要です

最初に、1日の必要条件を満たしたテーブルを作成します。

declare @stock TABLE ([StockDate] [datetime] NULL, 
         [stock] [int]) 

insert into @stock 
values ('2017/21/03', -3), ('2017/22/03', -1), 
     ('2017/23/03', -5), ('2017/24/03', 0), 
     ('2017/25/03', -4), ('2017/26/03', -7), 
     ('2017/27/03', 0); 

• 初期在庫：12
• 注文量：10

• 注文ポイント：3

• ため21/03/2017ストック= 12-3 = 9

• 22/03/2017在庫= 9-1 = 8
• 23/03/2017在庫= 8-5 = 3 < = OrderPoint（3 ） 次いでストック= 8-5 + 10 = 13

私が試み：

update s 
set s.stock = case 
       when s.StockDate = '2017/21/03' 
        then 12 
        else s2.stock 
       end + s.stock 
from @stock s 
left join @stock s2 on s2.StockDate = DATEADD(day, -1, s.StockDate) 

実際の結果：

Date     Stock 
2017-03-21 00:00:00.000 9 
2017-03-22 00:00:00.000 -4 
2017-03-23 00:00:00.000 -6 
2017-03-24 00:00:00.000 -5 
2017-03-25 00:00:00.000 -4 
2017-03-26 00:00:00.000 -11 
2017-03-27 00:00:00.000 -7 

まだ算出していない日-1 problemeとそのクエリで間違った結果を返します。

期待される結果：

---

Date     Stock 
2017-03-21 00:00:00.000 9 
2017-03-22 00:00:00.000 8 
2017-03-23 00:00:00.000 13 
2017-03-24 00:00:00.000 13 
2017-03-25 00:00:00.000 9 
2017-03-26 00:00:00.000 12 
2017-03-27 00:00:00.000 12 

"注文ポイントを" 調達する方法を提供した後、実行中の合計sum() over()を使用して、事前に

Answer 1

をありがとう：

declare @orderpoint table ([StockDate] [datetime] NULL, [stock] [int]); 
insert into @orderpoint values 
('20170323',10) 
,('20170326',10); 

select * 
    , RunningTotal = sum(stock) over (order by stockdate) 
from (
    select s.stockdate, stock = s.stock+isnull(o.stock,'') 
    from @stock s 
     left join @orderpoint o 
     on s.stockdate = o.stockdate 
    ) t 
order by stockdate 

rextester デモ：http://rextester.com/GHHU68033

リターン：

+------------+-------+--------------+ 
| stockdate | stock | RunningTotal | 
+------------+-------+--------------+ 
| 2017-03-20 | 12 |   12 | 
| 2017-03-21 | -3 |   9 | 
| 2017-03-22 | -1 |   8 | 
| 2017-03-23 |  5 |   13 | 
| 2017-03-24 |  0 |   13 | 
| 2017-03-25 | -4 |   9 | 
| 2017-03-26 |  3 |   12 | 
| 2017-03-27 |  0 |   12 | 
+------------+-------+--------------+ 

---

そしてupdateのために、あなたはcommon table expression

;with cte as (
    select 
     stockdate = convert(varchar(10),stockdate,120) 
    , stock 
    , RunningTotal = sum(stock) over (order by stockdate) 
    from (
     select s.stockdate, stock = s.stock+isnull(o.stock,'') 
     from @stock s 
     left join @orderpoint o 
      on s.stockdate = o.stockdate 
    ) t 
) 
update s 
    set stock = cte.stock 
from @stock s 
    inner join cte 
    on s.stockdate = cte.stockdate; 

select 
    stockdate = convert(varchar(10),stockdate,120) 
    , stock 
from @stock; 

リターンを使用することができます。

+------------+-------+ 
| stockdate | stock | 
+------------+-------+ 
| 2017-03-20 | 12 | 
| 2017-03-21 |  9 | 
| 2017-03-22 |  8 | 
| 2017-03-23 | 13 | 
| 2017-03-24 | 13 | 
| 2017-03-25 |  9 | 
| 2017-03-26 | 12 | 
| 2017-03-27 | 12 | 
+------------+-------+