使用日期和储罐名称的递归 T-SQL 查询
[英]Recursive T-SQL query Using Dates and Tank names
问题描述
好的，故事是这样的：我在一家将危险液体化学品回收成有用化学品的公司工作。 此过程的一部分涉及将这些材料从一个罐转移到另一个罐，因为回收过程的每个部分都已完成。 这些罐位于我们的罐区、蒸馏系统的一部分、轨道车或油罐车中。 我被分配的任务是编写一个查询，让经理可以拿起任何罐，并跟踪该材料在每个罐中的移动，直到它被处理或出售。 该数据库是在大约 10 年前建立的，比我还早，所以我只需要使用数据（即无法更改它以使其更易于使用）。
现在的表结构：适合这项工作的列是：
- 汤罐
- OpenDate（空罐开始接收材料的日期）
- CosedDate（储罐关闭、清空并准备好接受新材料的日期）
- DestinationTank（材料被移动到的罐）
- ProductName（我添加这个来验证动作。许多材料不能混合。有些是。如果算法不能正常工作，那些不能混合的可能会出现问题。）
所有储罐都有一个开放日期，但仍在填充或等待运出的储罐没有关闭日期。 大多数情况下，储罐是化学不可知的——例如，任何空储罐都可以用于任何化学品。 任何地方都可以预期 1 到 10 次移动。
我尝试使用递归 CTE T-SQL 算法，但在尝试将 SourceTank 与匹配的 DestinationTank 匹配时很快就迷路了，其中 SourceTank 的 CosedDate 介于 DestinationTank 的 OpenDate 和 CosedDate 之间，或者至少在或等于如果 DestinationTank 尚未关闭。
样本数据：
SourceTank StartDate ClosedDate DestinationTank ProductName
------------------------------------------------------------------------
TNK01-5 08/03/2017 08/10/2017 TNK30-6 Fuels
TNK01-5 08/07/2017 08/10/2017 TNK40-6 Fuels
TNK01-5 07/20/2017 07/31/2017 TNK01-5 Incin
TNK01-5 08/10/2017 08/17/2017 TNK30-6 Incin
TNK01-5 08/12/2017 08/17/2017 TNK30-6 Fuels
TNK03-5 08/13/2017 08/22/2017 TNK30-6 IBAC feed
TNK07-5 08/11/2017 08/17/2017 TNK40-6 Incin
TNK07-5 08/14/2017 08/29/2017 TNK40-6 Fuels
TNK07-5 07/15/2017 08/10/2017 TNK30-6 Picoline Cut
TNK07-5 08/03/2017 08/10/2017 TNK02-5 Pico 2nd Pass
TNK07-5 08/06/2017 08/17/2017 TNK40-6 Fuels
TNK08-5 08/05/2017 08/10/2017 TNK40-6 Fuels
TNK08-5 08/07/2017 08/08/2017 TNK30-6 Fuels
TNK08-5 08/10/2017 08/22/2017 TNK40-6 Water
TNK08-5 07/24/2017 08/10/2017 TNK02-5 Picoline Cut
TNK09-10 07/20/2017 NULL TNK30-6 Picoline Crude
TNK09-10 07/21/2017 08/04/2017 TNK30-6 Picoline Crude
TNK09-10 08/02/2017 08/10/2017 TNK30-6 Cyclo Waste
TNK09-10 08/05/2017 08/10/2017 TNK30-6 Cyclo Waste
TNK09-10 08/07/2017 08/10/2017 TNK30-6 Cyclo Waste
TNK09-10 08/04/2017 08/10/2017 TNK30-6 Cyclo Waste
TNK09-10 08/15/2017 08/22/2017 TNK30-6 IBAC feed
TNK09-10 08/11/2017 08/17/2017 TNK30-6 IBAC feed
TNK09-10 08/12/2017 08/17/2017 TNK30-6 IBAC feed
TNK30-6 08/08/2017 08/29/2017 TNK30-6 Cyclo Waste
TNK40-6 08/13/2017 08/22/2017 TNK30-6 IBAC:PRODUCT
TNK41-6 08/14/2017 09/27/2017 NATX25496 IBAC:PRODUCT
TNK51-12 07/26/2017 09/15/2017 TNK30-6 CYCLO Product
TNK62-12 07/28/2017 09/12/2017 TNK30-6 NON-RCRA NMP
TNK74-12 07/29/2017 NULL TNK30-6 Picoline Crude
TNK91-8 08/03/2017 08/22/2017 TNK08-5 Picoline Prod
这是我尝试过的。 不知道如何处理 NULL ClosingDate ..
WITH TanksCTE AS (
SELECT [TrackingNum]
,[TblTankTracking].[TankID]
,[StartDate]
,[TblTankTracking].[ProductID]
,[ClosedDate]
,[Destination]
FROM [MAFTS].[Laboratory].[TblTankTracking]
UNION ALL
SELECT TCTE.[TrackingNum]
,TRACK.[TankID]
,TRACK.[StartDate]
,TRACK.[ProductID]
,TRACK.[ClosedDate]
,TRACK.[Destination]
FROM TanksCTE AS TCTE INNER JOIN
[MAFTS].[Laboratory].[TblTankTracking] TRACK
ON TCTE.[TankID] = TRACK.[Destination]
AND TRACK.[ClosedDate] BETWEEN TCTE.[StartDate] AND TCTE.[ClosedDate]
)
SELECT * FROM TanksCTE
option (maxrecursion 0)
1 个解决方案
解决方案1
1 已采纳 2017-10-12 06:35:14
我不确定这会有多大用处，因为我不知道预期的结果应该是什么样子，但它可能会有所帮助。 我试图理解日期，这就是为什么你会看到它们连接成一个路径列。 我也不知道从哪里开始，所以我猜测每个坦克的最早开始日期。
MS SQL Server 2014 架构设置：
CREATE TABLE Table1
([SourceTank] varchar(8), [StartDate] datetime, [ClosedDate] datetime, [DestinationTank] varchar(9), [ProductName] varchar(14))
;
INSERT INTO Table1
([SourceTank], [StartDate], [ClosedDate], [DestinationTank], [ProductName])
VALUES
('TNK01-5', '2017-08-03 00:00:00', '08/10/2017', 'TNK30-6', 'Fuels'),
('TNK01-5', '2017-08-07 00:00:00', '08/10/2017', 'TNK40-6', 'Fuels'),
('TNK01-5', '2017-07-20 00:00:00', '07/31/2017', 'TNK01-5', 'Incin'),
('TNK01-5', '2017-08-10 00:00:00', '08/17/2017', 'TNK30-6', 'Incin'),
('TNK01-5', '2017-08-12 00:00:00', '08/17/2017', 'TNK30-6', 'Fuels'),
('TNK03-5', '2017-08-13 00:00:00', '08/22/2017', 'TNK30-6', 'IBAC feed'),
('TNK07-5', '2017-08-11 00:00:00', '08/17/2017', 'TNK40-6', 'Incin'),
('TNK07-5', '2017-08-14 00:00:00', '08/29/2017', 'TNK40-6', 'Fuels'),
('TNK07-5', '2017-07-15 00:00:00', '08/10/2017', 'TNK30-6', 'Picoline Cut'),
('TNK07-5', '2017-08-03 00:00:00', '08/10/2017', 'TNK02-5', 'Pico 2nd Pass'),
('TNK07-5', '2017-08-06 00:00:00', '08/17/2017', 'TNK40-6', 'Fuels'),
('TNK08-5', '2017-08-05 00:00:00', '08/10/2017', 'TNK40-6', 'Fuels'),
('TNK08-5', '2017-08-07 00:00:00', '08/08/2017', 'TNK30-6', 'Fuels'),
('TNK08-5', '2017-08-10 00:00:00', '08/22/2017', 'TNK40-6', 'Water'),
('TNK08-5', '2017-07-24 00:00:00', '08/10/2017', 'TNK02-5', 'Picoline Cut'),
('TNK09-10', '2017-07-20 00:00:00', NULL, 'TNK30-6', 'Picoline Crude'),
('TNK09-10', '2017-07-21 00:00:00', '08/04/2017', 'TNK30-6', 'Picoline Crude'),
('TNK09-10', '2017-08-02 00:00:00', '08/10/2017', 'TNK30-6', 'Cyclo Waste'),
('TNK09-10', '2017-08-05 00:00:00', '08/10/2017', 'TNK30-6', 'Cyclo Waste'),
('TNK09-10', '2017-08-07 00:00:00', '08/10/2017', 'TNK30-6', 'Cyclo Waste'),
('TNK09-10', '2017-08-04 00:00:00', '08/10/2017', 'TNK30-6', 'Cyclo Waste'),
('TNK09-10', '2017-08-15 00:00:00', '08/22/2017', 'TNK30-6', 'IBAC feed'),
('TNK09-10', '2017-08-11 00:00:00', '08/17/2017', 'TNK30-6', 'IBAC feed'),
('TNK09-10', '2017-08-12 00:00:00', '08/17/2017', 'TNK30-6', 'IBAC feed'),
('TNK30-6', '2017-08-08 00:00:00', '08/29/2017', 'TNK30-6', 'Cyclo Waste'),
('TNK40-6', '2017-08-13 00:00:00', '08/22/2017', 'TNK30-6', 'IBAC:PRODUCT'),
('TNK41-6', '2017-08-14 00:00:00', '09/27/2017', 'NATX25496', 'IBAC:PRODUCT'),
('TNK51-12', '2017-07-26 00:00:00', '09/15/2017', 'TNK30-6', 'CYCLO Product'),
('TNK62-12', '2017-07-28 00:00:00', '09/12/2017', 'TNK30-6', 'NON-RCRA NMP'),
('TNK74-12', '2017-07-29 00:00:00', NULL, 'TNK30-6', 'Picoline Crude'),
('TNK91-8', '2017-08-03 00:00:00', '08/22/2017', 'TNK08-5', 'Picoline Prod')
;
查询 1 ：
with TankStart as (
select
*
, row_number() over(partition by SourceTank order by StartDate ASC) as rn
from Table1
)
, TankPaths as (
SELECT
T.SourceTank
, T.DestinationTank
, T.StartDate
, T.ClosedDate
, T.ProductName
, CAST(T.DestinationTank AS varchar(max))
+ ' '
+ convert(varchar(10),T.ClosedDate,120)
AS Pathway
FROM TankStart T
WHERE T.rn = 1
union all
SELECT
T1.SourceTank
, T1.DestinationTank
, T1.StartDate
, T1.ClosedDate
, T1.ProductName
, M.Pathway
+ ' '
+ convert(varchar(10),T1.StartDate,120)
+ ', '
+ CAST(T1.DestinationTank AS varchar(max))
+ ', '
+ convert(varchar(10),T1.ClosedDate,120)
FROM Table1 T1
INNER JOIN TankPaths M ON M.DestinationTank = T1.SourceTank
AND M.ClosedDate <= T1.StartDate
where T1.SourceTank <> T1.DestinationTank
)
select *
from TankPaths
order by 1, 2, Pathway
结果：
| SourceTank | DestinationTank | StartDate | ClosedDate | ProductName | Pathway |
|------------|-----------------|----------------------|----------------------|----------------|------------------------------------------------------------------------------------|
| TNK01-5 | TNK01-5 | 2017-07-20T00:00:00Z | 2017-07-31T00:00:00Z | Incin | TNK01-5 2017-07-31 |
| TNK01-5 | TNK30-6 | 2017-08-03T00:00:00Z | 2017-08-10T00:00:00Z | Fuels | TNK01-5 2017-07-31 2017-08-03, TNK30-6, 2017-08-10 |
| TNK01-5 | TNK30-6 | 2017-08-10T00:00:00Z | 2017-08-17T00:00:00Z | Incin | TNK01-5 2017-07-31 2017-08-10, TNK30-6, 2017-08-17 |
| TNK01-5 | TNK30-6 | 2017-08-12T00:00:00Z | 2017-08-17T00:00:00Z | Fuels | TNK01-5 2017-07-31 2017-08-12, TNK30-6, 2017-08-17 |
| TNK01-5 | TNK40-6 | 2017-08-07T00:00:00Z | 2017-08-10T00:00:00Z | Fuels | TNK01-5 2017-07-31 2017-08-07, TNK40-6, 2017-08-10 |
| TNK03-5 | TNK30-6 | 2017-08-13T00:00:00Z | 2017-08-22T00:00:00Z | IBAC feed | TNK30-6 2017-08-22 |
| TNK07-5 | TNK30-6 | 2017-07-15T00:00:00Z | 2017-08-10T00:00:00Z | Picoline Cut | TNK30-6 2017-08-10 |
| TNK08-5 | TNK02-5 | 2017-07-24T00:00:00Z | 2017-08-10T00:00:00Z | Picoline Cut | TNK02-5 2017-08-10 |
| TNK09-10 | TNK30-6 | 2017-07-20T00:00:00Z | (null) | Picoline Crude | (null) |
| TNK30-6 | TNK30-6 | 2017-08-08T00:00:00Z | 2017-08-29T00:00:00Z | Cyclo Waste | TNK30-6 2017-08-29 |
| TNK40-6 | TNK30-6 | 2017-08-13T00:00:00Z | 2017-08-22T00:00:00Z | IBAC:PRODUCT | TNK01-5 2017-07-31 2017-08-07, TNK40-6, 2017-08-10 2017-08-13, TNK30-6, 2017-08-22 |
| TNK40-6 | TNK30-6 | 2017-08-13T00:00:00Z | 2017-08-22T00:00:00Z | IBAC:PRODUCT | TNK30-6 2017-08-22 |
| TNK41-6 | NATX25496 | 2017-08-14T00:00:00Z | 2017-09-27T00:00:00Z | IBAC:PRODUCT | NATX25496 2017-09-27 |
| TNK51-12 | TNK30-6 | 2017-07-26T00:00:00Z | 2017-09-15T00:00:00Z | CYCLO Product | TNK30-6 2017-09-15 |
| TNK62-12 | TNK30-6 | 2017-07-28T00:00:00Z | 2017-09-12T00:00:00Z | NON-RCRA NMP | TNK30-6 2017-09-12 |
| TNK74-12 | TNK30-6 | 2017-07-29T00:00:00Z | (null) | Picoline Crude | (null) |
| TNK91-8 | TNK08-5 | 2017-08-03T00:00:00Z | 2017-08-22T00:00:00Z | Picoline Prod | TNK08-5 2017-08-22 |
可以使用CTE或OVER将这种递归解决方案写到T-SQL查询中吗？
[英]Can this Recursive Solution be written up into a T-SQL Query using CTE or OVER?
声明:本站的技术帖子网页，遵循CC BY-SA 4.0协议，如果您需要转载，请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.