问题
I am using SQL Datawarehouse for my database
Find the V,T,C in sequence for a given detail group by Shipment_id order by TASK_SEQUENCE_NUMBER where the order of TASK_TYPE is V then T then C.
This value can be 0 if there is no values for V,T,C in sequence In case of V,T,C is more than once, we can sum the values and display
below is the table schema and data with the result expected which has to be incorporated in a stored procedure with more columns in the Task_Main Table.
CREATE TABLE [dbo].[Task_Detail](
[Task_Sid] [int] NULL, [Shipment_ID] [decimal](18, 0) NULL,
[TASK_SEQUENCE_No] [decimal](18, 0) NULL,
[TASK_TYPE] [varchar](1) NULL, [TASK_DURATION] [decimal](18, 0) NULL, [LOCATION_CODE] [varchar](15) NULL, [TaskStart] [bigint] NULL ) ON [PRIMARY]
CREATE TABLE [dbo].[Task_Header](
[Task_Sno] [int] NULL, [Shipment_ID] [decimal](18, 0) NULL,
[Vehicle_Id] [nchar](10) NULL, [DepotVisitStartTime] [datetime2](7) NULL, [Time_V] [bigint] NULL, [Time_T] [bigint] NULL, [Time_C] [bigint] NULL ) ON [PRIMARY]
INSERT [dbo].[Task_Detail] ([Task_Sid], [Shipment_ID], [TASK_SEQUENCE_No], [TASK_TYPE], [TASK_DURATION], [LOCATION_CODE], [TaskStart]) VALUES
(1, CAST(1 AS Decimal(18, 0)), CAST(1 AS Decimal(18, 0)), N'K', CAST(20 AS Decimal(18, 0)), N'LC39', 20),
(2, CAST(1 AS Decimal(18, 0)), CAST(7 AS Decimal(18, 0)), N'L', CAST(21 AS Decimal(18, 0)), N'10019514', 21),
(3, CAST(1 AS Decimal(18, 0)), CAST(3 AS Decimal(18, 0)), N'L', CAST(58 AS Decimal(18, 0)), N'LC38', 58),
(4, CAST(1 AS Decimal(18, 0)), CAST(6 AS Decimal(18, 0)), N'C', CAST(10 AS Decimal(18, 0)), NULL, 10),
(5, CAST(1 AS Decimal(18, 0)), CAST(8 AS Decimal(18, 0)), N'V', CAST(30 AS Decimal(18, 0)), N'10019514', 30),
(6, CAST(1 AS Decimal(18, 0)), CAST(10 AS Decimal(18, 0)), N'C', CAST(11 AS Decimal(18, 0)), N'RJC', 11),
(7, CAST(1 AS Decimal(18, 0)), CAST(9 AS Decimal(18, 0)), N'T', CAST(19 AS Decimal(18, 0)), N'10019027', 19)
, (8, CAST(2 AS Decimal(18, 0)), CAST(7 AS Decimal(18, 0)), N'V', CAST(29 AS Decimal(18, 0)), N'10018952', 29)
, (9, CAST(2 AS Decimal(18, 0)), CAST(4 AS Decimal(18, 0)), N'K', CAST(20 AS Decimal(18, 0)), NULL, 20)
, (10, CAST(2 AS Decimal(18, 0)), CAST(5 AS Decimal(18, 0)), N'V', CAST(0 AS Decimal(18, 0)), N'302', 0)
, (11, CAST(3 AS Decimal(18, 0)), CAST(12 AS Decimal(18, 0)), N'T', CAST(19 AS Decimal(18, 0)), N'10019514', 19)
, (12, CAST(3 AS Decimal(18, 0)), CAST(13 AS Decimal(18, 0)), N'E', CAST(11 AS Decimal(18, 0)), N'10019514', 11)
, (13, CAST(2 AS Decimal(18, 0)), CAST(3 AS Decimal(18, 0)), N'V', CAST(32 AS Decimal(18, 0)), N'', 32)
, (14, CAST(3 AS Decimal(18, 0)), CAST(4 AS Decimal(18, 0)), N'K', CAST(20 AS Decimal(18, 0)), N'10018952', 20)
, (15, CAST(3 AS Decimal(18, 0)), CAST(5 AS Decimal(18, 0)), N'V', CAST(0 AS Decimal(18, 0)), N'LC57', 0)
, (16, CAST(3 AS Decimal(18, 0)), CAST(8 AS Decimal(18, 0)), N'T', CAST(19 AS Decimal(18, 0)), N'10019514', 19)
, (17, CAST(3 AS Decimal(18, 0)), CAST(9 AS Decimal(18, 0)), N'C', CAST(10 AS Decimal(18, 0)), N'10019027', 10)
, (18, CAST(3 AS Decimal(18, 0)), CAST(3 AS Decimal(18, 0)), N'L', CAST(21 AS Decimal(18, 0)), N'LC38', 21)
, (19, CAST(3 AS Decimal(18, 0)), CAST(11 AS Decimal(18, 0)), N'V', CAST(30 AS Decimal(18, 0)), N'10019514', 30)
, (20, CAST(3 AS Decimal(18, 0)), CAST(10 AS Decimal(18, 0)), N'L', CAST(21 AS Decimal(18, 0)), N'10019514', 21)
, (21, CAST(3 AS Decimal(18, 0)), CAST(1 AS Decimal(18, 0)), N'T', CAST(19 AS Decimal(18, 0)), N'LC39', 19)
, (22, CAST(3 AS Decimal(18, 0)), CAST(2 AS Decimal(18, 0)), N'C', CAST(10 AS Decimal(18, 0)), N'302', 10)
, (23, CAST(1 AS Decimal(18, 0)), CAST(2 AS Decimal(18, 0)), N'V', CAST(0 AS Decimal(18, 0)), N'302', 0)
, (24, CAST(1 AS Decimal(18, 0)), CAST(4 AS Decimal(18, 0)), N'V', CAST(29 AS Decimal(18, 0)), N'10018952', 29)
, (25, CAST(1 AS Decimal(18, 0)), CAST(5 AS Decimal(18, 0)), N'T', CAST(19 AS Decimal(18, 0)), N'LC57', 19)
, (26, CAST(3 AS Decimal(18, 0)), CAST(6 AS Decimal(18, 0)), N'L', CAST(58 AS Decimal(18, 0)), N'10019514', 58)
, (27, CAST(3 AS Decimal(18, 0)), CAST(7 AS Decimal(18, 0)), N'V', CAST(29 AS Decimal(18, 0)), N'10019514', 29)
, (28, CAST(2 AS Decimal(18, 0)), CAST(6 AS Decimal(18, 0)), N'L', CAST(58 AS Decimal(18, 0)), N'10018952', 58)
, (29, CAST(2 AS Decimal(18, 0)), CAST(2 AS Decimal(18, 0)), N'L', CAST(35 AS Decimal(18, 0)), NULL, 35)
INSERT [dbo].[Task_Header] ([Task_Sno], [Shipment_ID], [Vehicle_Id], [DepotVisitStartTime], [Time_V], [Time_T], [Time_C]) VALUES
(1, CAST(1 AS Decimal(18, 0)), N'TN1 ', CAST(N'2019-02-15T07:25:33.0000000' AS DateTime2), NULL, NULL, NULL)
,(2, CAST(1 AS Decimal(18, 0)), N'TN1 ', CAST(N'2019-02-15T07:25:33.0000000' AS DateTime2), NULL, NULL, NULL)
,(3, CAST(1 AS Decimal(18, 0)), N'TN1 ', CAST(N'2019-02-15T07:25:33.0000000' AS DateTime2), NULL, NULL, NULL)
,(4, CAST(1 AS Decimal(18, 0)), N'TN1 ', CAST(N'2019-02-15T07:25:33.0000000' AS DateTime2), NULL, NULL, NULL)
,(5, CAST(2 AS Decimal(18, 0)), N'KA2 ', CAST(N'2019-02-15T06:12:52.0000000' AS DateTime2), NULL, NULL, NULL)
,(6, CAST(2 AS Decimal(18, 0)), N'KA2 ', CAST(N'2019-02-15T06:12:52.0000000' AS DateTime2), NULL, NULL, NULL)
,(7, CAST(3 AS Decimal(18, 0)), N'AP3 ', CAST(N'2019-02-15T06:32:52.0000000' AS DateTime2), NULL, NULL, NULL)
,(8, CAST(3 AS Decimal(18, 0)), N'AP3 ', CAST(N'2019-02-15T06:32:52.0000000' AS DateTime2), NULL, NULL, NULL)
,(9, CAST(3 AS Decimal(18, 0)), N'AP3 ', CAST(N'2019-02-15T13:12:21.0000000' AS DateTime2), NULL, NULL, NULL)
Expected Result with filter:
Sort Data by Shipment_ID , TASK_SEQUENCE_NUMBER filter where Location_code <> 'RJC' Join of TaskMaster and TaskDetails is Shipment_id TASK_TYPE should be V in a given sequence, next immediate with T , next immediate with C only considered for output
Sample data : result in column V : 59 is sum of "V" value for the given Shipment_Id result data is group by shipment_id
Expected Result
Sno Shipment_ID Vehicle ID DepotVisitStartTime V T C
1 1 TN1 2019-02-15 07:25:33.0000000 59 38 21
2 1 TN1 2019-02-15 07:25:33.0000000 59 38 21
3 1 TN1 2019-02-15 07:25:33.0000000 59 38 21
4 1 TN1 2019-02-15 07:25:33.0000000 59 38 21
5 2 KA2 2019-02-15 06:12:52.0000000 0 0 0
6 2 KA2 2019-02-15 06:12:52.0000000 0 0 0
7 3 AP3 2019-02-15 06:32:52.0000000 29 19 10
8 3 AP3 2019-02-15 06:32:52.0000000 29 19 10
9 3 AP3 2019-02-15 13:12:21.0000000 29 19 10
These columns are continuation of below post and i want to have an StoredProceudre to insert data into Task_Master table Aggregate Date and quantity in same table for given ID
回答1:
Code given by @Gordon seems to be mostly correct, just changing the column names in main logic, hope this helps!
select Shipment_ID,
sum(case when task_type = 'V' then task_duration else 0 end) as v_duration,
sum(case when task_type = 'T' then task_duration else 0 end) as t_duration,
sum(case when task_type = 'C' then task_duration else 0 end) as c_duration
from (select td.*,
lead(task_type) over (partition by Shipment_ID order by TASK_SEQUENCE_No) as next_task_type,
lead(task_type, 2) over (partition by Shipment_ID order by TASK_SEQUENCE_No) as next2_task_type,
lag(task_type) over (partition by Shipment_ID order by TASK_SEQUENCE_No) as prev_task_type,
lag(task_type, 2) over (partition by Shipment_ID order by TASK_SEQUENCE_No) as prev2_type
from task_detail td --WHERE Location_Code <>'RJC'
) td
where (task_type = 'V' and next_task_type = 'T' and next2_task_type = 'C') or
(prev_task_type = 'V' and task_type = 'T' and next_task_type = 'C') or
(prev2_type = 'V' and prev_task_type = 'T' and task_type = 'C')
group by Shipment_ID;
Adding the image of the result obtained.Results
回答2:
You can use lead() and lag():
select ship_id,
sum(case when task_type = 'V' then task_duration else 0 end) as v_duration,
sum(case when task_type = 'T' then task_duration else 0 end) as t_duration,
sum(case when task_type = 'C' then task_duration else 0 end) as c_duration
from (select td.*,
lead(task_type) over (partition by ship_id order by task_sid) as next_task_type,
lead(task_type, 2) over (partition by ship_id order by task_sid) as next2_task_type,
lag(task_type) over (partition by ship_id order by task_sid) as prev_task_type,
lag(task_type, 2) over (partition by ship_id order by task_sid) as prev2_type,
from task_detail td
) td
where (task_type = 'V' and next_task_type = 'T' and next2_task_type = 'C') or
(prev_task_type = 'V' and task_type = 'T' and next_task_type = 'C') or
(prev2_task_type = 'V' and prev_task_type = 'T' and task_type = 'C')
group by ship_id;
来源：https://stackoverflow.com/questions/54831602/value-based-on-sequence-for-each-detail