Group By with 2 parameters and datetime and add next dateTime.

Question

Group By with 2 parameters and datetime and add next dateTime.

onipif

SSC Enthusiast

Points: 103
More actions
January 24, 2009 at 10:28 am

#383799

Hello everybody,
In first, this is my first topic on your forum(I'm french), but I read more (interesting) articles, this is a very good SQL website and forum 😉
My probem :
Since last Week, I have a problem with data and I don't try to resolve it, but when I read your newsletter and the "Linking to the Previous Row" I thinks this is the good way, so I hope that you can help me, because I don't try it by myself 😉
(I'm on SQLServer2000, not 2005 so I cannot use the CTE.)
An exemple of data :
(I have add "id" column but only to help comprehension, in my process table, I have only dT as primary key, thus I cannot order by Id)
--===== If the test table already exists, drop it
IF OBJECT_ID('TempDB..#mytableDamien','U') IS NOT NULL
DROP TABLE #mytableDamien
--===== Create the test table with
CREATE TABLE #mytableDamien
(
id INT IDENTITY(1,1) UNIQUE,
dT DATETIME PRIMARY KEY CLUSTERED,
param1 VARCHAR(8),
param2 VARCHAR(8)
)
--===== Setup any special required conditions especially where dates are concerned
SET DATEFORMAT YMD
--===== All Inserts into the IDENTITY column
SET IDENTITY_INSERT #mytableDamien ON
INSERT INTO #mytableDamien
(id, dT, param1, param2)
SELECT '1','2009-01-13 18:46:02.000','p1_5','1' UNION ALL
SELECT '2','2009-01-13 16:07:38.000','p1_5','1' UNION ALL
SELECT '3','2009-01-12 16:17:23.000','p1_3','1' UNION ALL
SELECT '4','2009-01-12 13:34:07.000','p1_3','1' UNION ALL
SELECT '5','2009-01-12 12:49:31.000','p1_2','1' UNION ALL
SELECT '6','2009-01-11 18:52:09.000','p1_5','1' UNION ALL
SELECT '7','2009-01-11 18:51:49.000','p1_5','2' UNION ALL
SELECT '8','2009-01-11 17:54:51.000','p1_5','2' UNION ALL
SELECT '9','2009-01-11 14:57:09.000','p1_3','2' UNION ALL
SELECT '10','2009-01-10 14:25:05.000','p1_3','2' UNION ALL
SELECT '11','2009-01-10 14:14:59.000','p1_1','2' UNION ALL
SELECT '12','2009-01-08 22:08:26.000','p1_5','1'
SET IDENTITY_INSERT #mytableDamien OFF
--===== Data selection
SELECT id, dT, param1, param2 FROM #mytableDamien ORDER BY dT DESC
-id---Dt------------------------Param1---Param2
-01--2009-01-13 18:46:02.000---p1_5-----1
-02--2009-01-13 16:07:38.000---p1_5-----1
-03--2009-01-12 16:17:23.000---p1_3-----1
-04--2009-01-12 13:34:07.000---p1_3-----1
-05--2009-01-12 12:49:31.000---p1_2-----1
-06--2009-01-11 18:52:09.000---p1_5-----1
-07--2009-01-11 18:51:49.000---p1_5-----2
-08--2009-01-11 17:54:51.000---p1_5-----2
-09--2009-01-11 14:57:09.000---p1_3-----2
-10--2009-01-10 14:25:05.000---p1_3-----2
-11--2009-01-10 14:14:59.000---p1_1-----2
-12--2009-01-08 22:08:26.000---p1_5-----1
I want to GROUP BY param1, Param2, but only when an other couple of Param1, Param2 exists between it. So, the Dt of result will be the start/first date time. There is no sequentiel logic on the Dt, it's possible to have only few minuts between 2 rows, or X days...
The result of my request will be :
-id---Dt------------------------Param1---Param2
-02--2009-01-13 16:07:38.000---p1_5-----1
-04--2009-01-12 13:34:07.000---p1_3-----1
-05--2009-01-12 12:49:31.000---p1_2-----1
-06--2009-01-11 18:52:09.000---p1_5-----1
-08--2009-01-11 17:54:51.000---p1_5-----2
-10--2009-01-10 14:25:05.000---p1_3-----2
-11--2009-01-10 14:14:59.000---p1_1-----2
-12--2009-01-08 22:08:26.000---p1_5-----1
Moreover, I want to add the end date time witch will be the Dt of the first next row of group by Param1, Param2 and order by Dt DESC.
So the result will be :
-id---startDt--------------------endDt-----------------------Param1---Param2
-02--2009-01-13 16:07:38.000---GETDATE()-----------------p1_5-----1
-04--2009-01-12 13:34:07.000---2009-01-13 16:07:38.000---p1_3-----1
-05--2009-01-12 12:49:31.000---2009-01-12 13:34:07.000---p1_2-----1
-06--2009-01-11 18:52:09.000---2009-01-12 12:49:31.000---p1_5-----1
-08--2009-01-11 17:54:51.000---2009-01-11 18:52:09.000---p1_5-----2
-10--2009-01-10 14:25:05.000---2009-01-11 17:54:51.000---p1_3-----2
-11--2009-01-10 14:14:59.000---2009-01-10 14:25:05.000---p1_1-----2
-12--2009-01-08 22:08:26.000---2009-01-10 14:14:59.000---p1_5-----1
An example would be the appropriate measurements using equipment from a not fixed time(minuts, hours or days) on 2 characteristics (Param1 and Param2) of a product, and I would like to response at question : "Was type of products was manufactured between a given time (Between dT) and the order of production with the period (startdT and endDt) ?
I hope that you will be help me because I don't find the good way by myself,
Best regards,
Damien.
PS :
I hope that I have a "good english" and that you understand me as well 😉

Viewing 3 posts - 1 through 2 (of 2 total)

You must be logged in to reply to this topic. Login to reply

Jeff Moden SSC Guru Points: 1004323 More actions · Answer 1

I believe that the reason you're not getting any bites on this problem is that most people don't want to work so hard trying to test your data. Please see the link in my signature below for the best way to post readily consumable data.

Also, people like to know WHY you need to do something like this. And just saying "It's a requirement" doesn't quite get it. People like to know the business reasons for doing such things because, a lot of times, there's better ways to do things than the way you may be thinking.

And, welcome aboard.

--Jeff Moden

RBAR is pronounced "ree-bar" and is a "Modenism" for Row-By-Agonizing-Row.
First step towards the paradigm shift of writing Set Based code:
________Stop thinking about what you want to do to a ROW... think, instead, of what you want to do to a COLUMN.

Change is inevitable... Change for the better is not.

Helpful Links:
How to post code problems
How to Post Performance Problems
Create a Tally Function (fnTally)

onipif SSC Enthusiast Points: 103 More actions · Answer 2

Hello RBAR,

Thanks for your answer and for your link about how post best data/code.

I have edit my first post according it.