Bcn78
asked on
Access Query
I'm struggling with a query in access. I'm using the DSUM function to calculate a cumulative total for COGS that starts over each year. The issue is that I need the cumulative total to not start over each year. I'm using the design view in access to attempt the query, but have copied the sql below (as this may be most helpful for the experts). So I need a cumulative total by Acctnbr and by Class that doesn't restart annually. Right now I can get it to a cumulative total by Acctnbr and by Class that restarts annually. I've exhausted my capabilities and am hoping for some good suggestions. Thanks.
SELECT Perpost, Year, Month, LocationCode, LocationDescription, LocationName, Class, State, LocationType, BenchmarkFacilityLocationC
FROM [tblUploadOccupancy(All)Fl
GROUP BY Perpost, Year, Month, LocationCode, LocationDescription, LocationName, Class, State, LocationType, BenchmarkFacilityLocationC
HAVING (((COGSFlag)="COGS"))
ORDER BY AcctNbr;
SELECT Perpost, Year, Month, LocationCode, LocationDescription, LocationName, Class, State, LocationType, BenchmarkFacilityLocationC
FROM [tblUploadOccupancy(All)Fl
GROUP BY Perpost, Year, Month, LocationCode, LocationDescription, LocationName, Class, State, LocationType, BenchmarkFacilityLocationC
HAVING (((COGSFlag)="COGS"))
ORDER BY AcctNbr;
ASKER CERTIFIED SOLUTION
membership
Create a free account to see this answer
Signing up is free and takes 30 seconds. No credit card required.
Also note that I changed the "HAVING" clause to a "WHERE" clause. I did this to limit the number of records that get considered in the main query. If you have the criteria in a WHERE clause, it excludes records which don't meet the criteria before the majority of the processing. If you put it in a HAVING clause, the query will process the SUM for all of the [COGSFlag] values and then exclude those that don't meet your final criteria from the final results.
To do this in design view, you must add a separate column in the query in the Totals row of the grid, select "WHERE" instead of "Group By"
BTW, if you have more than one record for each AcctNbr, Class, Year, Month combination in your table, then you may need to sum the [Amount] column as well, to get the sum of [Amount] for each month.
One last thing, [Year] and [Month] are reserved word and really should not be used as field names. You should generally give your fields meaningful names, so these might be better named like Transaction_Year and Transaction_Month, or something like that. You should also be advised that if you persist in using these field names, you should always wrap them in brackets [ ] to prevent confusion with the functions associated with those names.
Dale
To do this in design view, you must add a separate column in the query in the Totals row of the grid, select "WHERE" instead of "Group By"
BTW, if you have more than one record for each AcctNbr, Class, Year, Month combination in your table, then you may need to sum the [Amount] column as well, to get the sum of [Amount] for each month.
One last thing, [Year] and [Month] are reserved word and really should not be used as field names. You should generally give your fields meaningful names, so these might be better named like Transaction_Year and Transaction_Month, or something like that. You should also be advised that if you persist in using these field names, you should always wrap them in brackets [ ] to prevent confusion with the functions associated with those names.
Dale
ASKER
Thanks Dale. I appreciate your help. Understanding your solution is little beyond my current capability, so it's very helpful for me to study and understand your response to help advance my skill set. I'm going to work with this for a while and may circle back with a question or two if you don't mind. Thanks again.
brett
brett
ASKER
Separately, I am using this to make a table that I can use in a subsequent query to forecast WIP and finished goods inventory
ASKER
This is great. I really appreciate it. The below was very helpful as well, it made the query run 10x faster. Thanks again Dale.
-Brett
Also note that I changed the "HAVING" clause to a "WHERE" clause. I did this to limit the number of records that get considered in the main query. If you have the criteria in a WHERE clause, it excludes records which don't meet the criteria before the majority of the processing. If you put it in a HAVING clause, the query will process the SUM for all of the [COGSFlag] values and then exclude those that don't meet your final criteria from the final results.
-Brett
Also note that I changed the "HAVING" clause to a "WHERE" clause. I did this to limit the number of records that get considered in the main query. If you have the criteria in a WHERE clause, it excludes records which don't meet the criteria before the majority of the processing. If you put it in a HAVING clause, the query will process the SUM for all of the [COGSFlag] values and then exclude those that don't meet your final criteria from the final results.
glad to help, Brett.
non-equi JOINs can be confusing, but can generally be re-written as a WHERE clause as below:
SELECT T1.Perpost, T1.[Year], T1.[Month], T1.LocationCode, T1.LocationDescription, T1.LocationName, T1.Class, T1.State, T1.LocationType, T1.BenchmarkFacilityLocati
FROM [tblUploadOccupancy(All)Fl
INNER JOIN [tblUploadOccupancy(All)Fl
ON T1.[AcctNbr] = T2.[AcctNbr]
AND T1.[Class] = T2.[Class]
WHERE (T1.COGSFlag="COGS")
AND (DateSerial(T1.[Year], T1.[Month] +1, 0) >= DateSerial(T2.[Year], T2.[Month] + 1, 0))
GROUP BY T1.Perpost, T1.[Year], T1.[Month], T1.LocationCode, T1.LocationDescription, T1.LocationName, T1.Class, T1.State, T1.LocationType, T1.BenchmarkFacilityLocati
ORDER BY T1.AcctNbr, T1.Year, T1.Month
In this version, which you could edit in the query design grid, I've simply moved the non-equi JOIN and moved that clause into the WHERE clause. Consider a simple example of a table (tblItemSales) with sales for a single item and fields (ItemID, SalesDate, Quantity) and just a couple of records:
ItemID SalesDate Quantity
1 7/1/16 100
1 7/2/16 150
1 7/3/16 125
if you create a query:
SELECT T1.ItemID, T1.SalesDate, T1.Quantity, T2.SalesDate, T2.Quantity
FROM tblItemSales as T1
INNER JOIN tblItemSales as T2 ON T1.ItemID = T2.ItemID
you would get 9 records (three records from T2 for each record in T1).
If you added a WHERE clause to only include the record from T2 which are <= T1.Sales date then you would initially get 9 records in the recordset, but the WHERE clause would filter out three, leaving 6 records (1 from T2 for T1.SalesDate = 7/1/16, 2 records from T2 for T1.SalesDate = 7/2/16, and 3 records from T2 for T1.SalesDate = 7/3/16).
Theoretically, if you move that WHERE clause into a non-equi JOIN, you would only get the 6 records initially. So, the non-equi JOIN would get reduce the number of records to be processed.
Hope this helps.
Dale
non-equi JOINs can be confusing, but can generally be re-written as a WHERE clause as below:
SELECT T1.Perpost, T1.[Year], T1.[Month], T1.LocationCode, T1.LocationDescription, T1.LocationName, T1.Class, T1.State, T1.LocationType, T1.BenchmarkFacilityLocati
FROM [tblUploadOccupancy(All)Fl
INNER JOIN [tblUploadOccupancy(All)Fl
ON T1.[AcctNbr] = T2.[AcctNbr]
AND T1.[Class] = T2.[Class]
WHERE (T1.COGSFlag="COGS")
AND (DateSerial(T1.[Year], T1.[Month] +1, 0) >= DateSerial(T2.[Year], T2.[Month] + 1, 0))
GROUP BY T1.Perpost, T1.[Year], T1.[Month], T1.LocationCode, T1.LocationDescription, T1.LocationName, T1.Class, T1.State, T1.LocationType, T1.BenchmarkFacilityLocati
ORDER BY T1.AcctNbr, T1.Year, T1.Month
In this version, which you could edit in the query design grid, I've simply moved the non-equi JOIN and moved that clause into the WHERE clause. Consider a simple example of a table (tblItemSales) with sales for a single item and fields (ItemID, SalesDate, Quantity) and just a couple of records:
ItemID SalesDate Quantity
1 7/1/16 100
1 7/2/16 150
1 7/3/16 125
if you create a query:
SELECT T1.ItemID, T1.SalesDate, T1.Quantity, T2.SalesDate, T2.Quantity
FROM tblItemSales as T1
INNER JOIN tblItemSales as T2 ON T1.ItemID = T2.ItemID
you would get 9 records (three records from T2 for each record in T1).
If you added a WHERE clause to only include the record from T2 which are <= T1.Sales date then you would initially get 9 records in the recordset, but the WHERE clause would filter out three, leaving 6 records (1 from T2 for T1.SalesDate = 7/1/16, 2 records from T2 for T1.SalesDate = 7/2/16, and 3 records from T2 for T1.SalesDate = 7/3/16).
Theoretically, if you move that WHERE clause into a non-equi JOIN, you would only get the 6 records initially. So, the non-equi JOIN would get reduce the number of records to be processed.
Hope this helps.
Dale
Open in new window
with thisOpen in new window
It might help if you posted some representative sample of the data. I'm not sure what you mean when you write: