asked on

Using a table of dates in a query - why?

I inherited this query from my predecessor. It returns the correct data, but it seems like a hard way to get there. He has a table that contains all the dates between 1-1-2000 and 12-31-2030 which he uses extensively in queries. It seems like there has to be a more elegant way of doing this, but I'm far from being an expert so maybe there's a reason for going this route. It also seems like there should be a better way of getting the count of late deliveries (Line 79). We don't care how late they are, we just need to know the number of late deliveries to get the percentage of deliveries that were not late, broken out between the two warehouses. Below is the query with a small piece of sample data. Normally this would cover one or more whole months.

DECLARE @Calendar TABLE

(
	CalendarDate DATE
)

DECLARE @PorMasterDetail TABLE
(
     MLatestDueDate  DATETIME,
     MLastReceiptDat DATETIME,
	 MWarehouse CHAR(2)
)

INSERT INTO @PorMasterDetail
VALUES ('2014-12-04 00:00:00.000', '2014-12-03 00:00:00.000', '20'),
('2014-12-05 00:00:00.000', '2014-12-04 00:00:00.000', '10'),
('2014-12-08 00:00:00.000', '2014-12-02 00:00:00.000', '10'),
('2014-12-08 00:00:00.000', '2014-12-02 00:00:00.000', '10'),
('2014-12-05 00:00:00.000', '2014-12-03 00:00:00.000', '20'),
('2014-12-15 00:00:00.000', '2014-12-03 00:00:00.000', '20'),
('2014-12-01 00:00:00.000', '2014-12-01 00:00:00.000', '10'),
('2014-12-03 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-09 00:00:00.000', '2014-12-04 00:00:00.000', '20'),
('2014-12-05 00:00:00.000', '2014-12-05 00:00:00.000', '10'),
('2014-12-05 00:00:00.000', '2014-12-04 00:00:00.000', '10'),
('2015-04-04 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-05 00:00:00.000', '2014-12-05 00:00:00.000', '10'),
('2014-12-02 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-02 00:00:00.000', '2014-12-03 00:00:00.000', '20'),
('2014-12-02 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-02 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-02 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-02 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-10 00:00:00.000', '2014-12-05 00:00:00.000', '10'),
('2014-12-15 00:00:00.000', '2014-12-04 00:00:00.000', '20'),
('2014-12-15 00:00:00.000', '2014-12-05 00:00:00.000', '20'),
('2014-12-05 00:00:00.000', '2014-12-04 00:00:00.000', '20'),
('2014-12-10 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-10 00:00:00.000', '2014-12-01 00:00:00.000', '20'),
('2014-12-05 00:00:00.000', '2014-12-01 00:00:00.000', '10'),
('2014-12-04 00:00:00.000', '2014-12-01 00:00:00.000', '20'),
('2014-12-04 00:00:00.000', '2014-12-01 00:00:00.000', '20'),
('2014-12-02 00:00:00.000', '2014-12-03 00:00:00.000', '20'),
('2014-12-18 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-03 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-03 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-02 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-02 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-02 00:00:00.000', '2014-12-02 00:00:00.000', '20'),
('2014-12-05 00:00:00.000', '2014-12-05 00:00:00.000', '10'),
('2014-12-01 00:00:00.000', '2014-12-01 00:00:00.000', '10')

INSERT INTO @Calendar
VALUES ('2014-12-01'),
('2014-12-02'),
('2014-12-03'),
('2014-12-04'),
('2014-12-05')

SELECT
    C.[Year],
    C.[Month],
    t1.Hardware,
    t1.[Raw Materials]
FROM (SELECT DISTINCT
		YEAR(C.CalendarDate) AS [Year],
		MONTH(C.CalendarDate) AS M,
        DATENAME(m, C.CalendarDate) AS [Month]
        FROM   @Calendar C
        WHERE  C.CalendarDate >= '20141201' AND C.CalendarDate <= '20141212') C
       LEFT JOIN (SELECT
					W10.[Year],
                    W10.[M],
                    W20.Hardware,
                    W10.[Raw Materials]
                  FROM   (SELECT
							CAST(YEAR([MLastReceiptDat]) AS INT) AS [Year],
                            CAST(MONTH([MLastReceiptDat]) AS INT) AS [M],
                            CAST(CAST((COUNT(*) - SUM(CASE WHEN MLastReceiptDat - MLatestDueDate > 0 THEN 1 ELSE 0
                                                      END))AS DECIMAL(10, 4)) / CAST(COUNT(*) AS DECIMAL(10, 4))AS DECIMAL(10, 4)) AS [Hardware]
                          FROM @PorMasterDetail
                          WHERE  (MWarehouse = '20')
                          GROUP  BY YEAR([MLastReceiptDat]), MONTH([MLastReceiptDat]), MWarehouse) W20
                          INNER JOIN (SELECT
										CAST(YEAR([MLastReceiptDat]) AS INT) AS [Year],
                                        CAST(MONTH([MLastReceiptDat]) AS INT) AS [M],
                                        CAST(CAST((COUNT(*) - SUM(CASE WHEN MLastReceiptDat - MLatestDueDate > 0 THEN 1 ELSE 0
                                                                  END))AS DECIMAL(10, 4)) / CAST(COUNT(*) AS DECIMAL(10, 4)) AS DECIMAL(10, 4)) AS [Raw Materials]
                                     FROM   @PorMasterDetail t2
                                     WHERE  (MWarehouse = '10')
                                     GROUP  BY YEAR([MLastReceiptDat]), MONTH([MLastReceiptDat]), MWarehouse) W10
									 ON W10.[Year] = W20.[Year] AND W10.[M] = W20.[M]) t1
									 ON t1.[Year] = C.[Year] AND t1.[M] = C.[M]
ORDER  BY C.[Year], C.[M]

Open in new window

ASKER CERTIFIED SOLUTION

Steve Wales

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

Jim Horn

Here's my article that explains it ... Build your own calendar table to perform complex date expressions

The short answer: Not everybody lives in your world, and SQL Server isn't built to handle everybody's different world when it comes to dates.

Some places have goofy holidays of their own where stores are closed.

The entire state of Wisconsin shuts down for Green Bay Packers football games. Businesses have to plan for that.

My last client was an airline, and I had to build functionality to accurately the predict the effects of Daylight Savings Time (or lack thereof) for all airports.

Scott Pletcher

It is just silly to use a calendar table, 365 row summary, to get 12 individual months that are known up front. Many people get so wedded to the notion of a calendar table they can't ignore it even when it makes no sense for a specific query. [Btw, how far then do you take this? Some companies close for half-days or even hours, so do we really need a table with every hour between 2000 and 2099?] A nonwork days calendar makes perfect sense, but using a full calendar for basic date actions is not justified to me.

As to the rest, the @PorMasterDetail table needs to be read only once, not twice. And the dates should not be split into separate yr & mth until the very end, as that is effectively just formatting of the underlying date.

Zberteoc

I prefer this solution:

USE YourDatabase
GO

SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
CREATE FUNCTION [dbo].[fnTally]()
RETURNS TABLE --WITH SCHEMABINDING 
AS
/*******************************************************************************\
Function	: fnTally

Purpose		: returns a set with numbers from 1 to 10,000 
			  to be used in parsing and sequential data generation whithout loop
			  
Parameters	: no parameters

Invoke		:
	
		select * from [dbo].[fnTally]()
		select N from [dbo].[fnTally]()
		select substring('abcdef',N,1) as chr from [dbo].[fnTally]() where N<len('abcdef') -- parsing a string
		select dateadd(dd, N, '2007-01-01') as dte from [dbo].[fnTally]() --gets dates for about 30 years

\*******************************************************************************/
RETURN
	WITH 
	E1(N) AS 
	( --10E+1 or 10 rows
		 SELECT 1 UNION ALL SELECT 1 UNION ALL SELECT 1 UNION ALL
		 SELECT 1 UNION ALL SELECT 1 UNION ALL SELECT 1 UNION ALL
		 SELECT 1 UNION ALL SELECT 1 UNION ALL SELECT 1 UNION ALL SELECT 1
	),                         
   E2(N) AS 
   ( --10E+2 or 100 rows	
		SELECT 1 FROM E1 a, E1 b
	),
   E4(N) AS 
   ( --10E+4 or 10,000 rows max
		SELECT 1 FROM E2 a, E2 b
	)
			 SELECT ROW_NUMBER() OVER (ORDER BY (SELECT NULL)) as N FROM E4
	;

Open in new window

The function itself generates numbers from 1 to 10000 and, as specified in commented examples inside the function, you can use it to produce dates for 30 years like this:

select dateadd(dd, N, '2007-01-01') as dte from [dbo].[fnTally]()

Open in new window

SOLUTION

Vitor Montalvão

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

BAM

ASKER

Thanks to everyone for your helpful comments. All are appreciated. I see a should not have asked two questions in the same topic. Live and learn.

Steve: I watched the video and it makes sense. The table I currently have contains a single date, so I wasn't seeing much benefit in that approach.

Vitor: Thank you so much for straightening this out.

Jim: Your article has been removed.

Jim Horn

Oops. I posted the link to the draft, which was deleted when I published the final version. My bad.