Link to home
Start Free TrialLog in
Avatar of James Elliott
James ElliottFlag for United Kingdom of Great Britain and Northern Ireland

asked on

Stored Procedure - Track movements in queue


I'm making a complete meal out of a stored procedure that I think should be very simple. Please help.

I have Table A which holds financial information in queue, which I snapshot throughout the day on different days (Uploaded_DT).

I have a Query B which gives me a list of the days that the snapshots were taken, and gives me the last snapshot time of the day.

I join this to Table A to give me 1 lot of data for each day, taken from the last snapshot of the day.

In Table A, I have [Date Created] and [Invoice Total] from which I want the stored procedure to aggregate and return the SUM(Invoice Total) for the first day, less the SUM(Invoice Total) for the final day, plus any SUM(Invoice Values) with a [Date Created] in the period.

Here's where I've got to.



ALTER procedure [dbo].[spGetFinancialMovementsBetweenDates]
	@dtStart datetime,
	@dtEnd datetime

DECLARE @dOpeningBalance as DECIMAL(10,2)
DECLARE @dClosingBalance as DECIMAL(10,2)
DECLARE @dCreatedInPeriod as DECIMAL(10,2) 

		SELECT c.dt, SUM(c.InvTotal) as bf, SUM(c.CreatedInDT) as created_dt INTO #tt
					a.Uploaded_DT, b.dt, MAX(a.[Invoice Total]) as InvTotal, CASE WHEN CONVERT(varchar(10),MAX(a.[Date Created]),103) = b.dt THEN MAX(a.[Invoice Total]) ELSE 0 END as CreatedInDT
					TableA a
						QueryB b
					ON b.uploaded_dt = a.Uploaded_DT
					AND a.Uploaded_DT >= @dtStart
					AND a.Uploaded_DT <= DATEADD(day,1,@dtEnd)

					WHERE a.[Date Created] IS NOT NULL

					GROUP BY a.uploaded_dt, b.dt, a.[Payee Number], a.[Invoice Number]
				  ) c
            GROUP BY c.dt

SET @dOpeningBalance = (SELECT bf FROM #tt  WHERE dt = CONVERT(varchar(10),@dtStart,103))
SET @dClosingBalance = (SELECT bf FROM #tt WHERE dt = CONVERT(varchar(10),@dtEnd,103))
SET @dCreatedInPeriod = (SELECT SUM(created_dt), CASE WHEN dt > CONVERT(varchar(10),@dtStart,103) and dt <= CONVERT(varchar(10),@dtEnd,103) THEN 1 ELSE 0 END as grp FROM #tt GROUP BY CASE WHEN dt > CONVERT(varchar(10),@dtStart,103) and dt <= CONVERT(varchar(10),@dtEnd,103) THEN 1 ELSE 0 END )

SELECT @dOpeningBalance as bf, @dCreatedInPeriod as [in], (@dopeningbalance - @dClosingBalance) + @dCreatedInPeriod as [out], @dClosingBalance as cf



Open in new window

Avatar of Vitor Montalvão
Vitor Montalvão
Flag of Switzerland image

Can you post a sample data? And which data are returning by your SP and what should returns so we can see what's missing in your query.

Also I can't understand the use of the CASE in the subselect.
And can you explain your data model? Cause the task sounds like it is part of an ETL process.
Avatar of James Elliott


Ste5an => I've googled ETL :) I don't know whether it applies or not! My setup doesn't sound like a classical ETL process. Can you explain what you want this information for, and I'll try and explain the data model in the most useful way possible. I'm not particularly well versed in this area.

Vitor => I'm fairly restricted in what I can share as regards data, but please see attached an example.

Thanks in advance.
It's the way a data warehouse gets its data: extract, transform, load. "Snapshots", "aggregates" really sound like a DW issue. This can mean that a different model is more approriate.
Avatar of Vitor Montalvão
Vitor Montalvão
Flag of Switzerland image

Link to home
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Thanks Vitor.

Nearly there I think.

I'm struggling with a.same_day though. This isn't in my source data.

James, same_day it's in the Excel file you posted here. I thought those columns are fields of your table.
ah ok, no, some of the columns were calculated on the spreadsheet itself to demonstrate what I needed.

I think you've given me enough to go on though with the rollup, and the method of selecting TOP 1 from a sorted listed. I'm going to try and work this out, so I'll post another question to take me further if I need it.

Great direction. Thanks