Click here to monitor SSC
SQLServerCentral is supported by Red Gate Software Ltd.
 
Log in  ::  Register  ::  Not logged in
 
 
 
        
Home       Members    Calendar    Who's On


Add to briefcase

Can a JOIN be used here instead of a UNION? Expand / Collapse
Author
Message
Posted Wednesday, September 3, 2014 12:45 AM
Valued Member

Valued MemberValued MemberValued MemberValued MemberValued MemberValued MemberValued MemberValued Member

Group: General Forum Members
Last Login: Sunday, December 14, 2014 1:08 PM
Points: 60, Visits: 126
Hi Everyone,

I am in the midst of writing a report which shows the 'Total Sales', 'Gross Profit', and 'Gross Profit %' by Sales Person. Traditionally I have used a UNION between two result sets to achieve this task. However I am wondering it if is possible to achieve the same outcome with a JOIN as shown below?

SELECT
T0.SlpCode
, T0.SalesPerson
, T0.Branch
, (T0.TotalSales - T1.TotalSales) AS 'Total Sales'
, ((T0.TotalSales - T1.TotalSales) - (T0.StockValue + T1.StockValue)) AS 'Gross Profit'
, CAST(((T0.TotalSales - T1.TotalSales) - (T0.StockValue + T1.StockValue)) / NULLIF((T0.StockValue + T1.StockValue), 0) * 100 AS decimal(15,2)) AS 'Gross Profit %'

FROM
(
/*** SAP - Invoices by Sales Person ***/

SELECT
T1.SlpCode
, T2.SlpName AS 'SalesPerson'
, T2.U_INE_Branch AS 'Branch'
, CAST(SUM(T0.LineTotal) AS decimal(15,2)) AS 'TotalSales'
, CAST(SUM(T0.StockValue) AS decimal(15,2)) AS 'StockValue'

FROM AU.dbo.INV1 T0
INNER JOIN AU.dbo.OINV T1 ON T1.DocEntry = T0.DocEntry
LEFT JOIN AU.dbo.OSLP T2 ON T2.SlpCode = T1.SlpCode

WHERE T1.DocDate >= '2014-08-01' AND T1.DocDate <= '2014-09-01'

GROUP BY T1.SlpCode, T2.SlpName, T2.U_INE_Branch
) AS T0
FULL JOIN
(
/*** SAP - Credits by Sales Person ***/

SELECT
T1.SlpCode
, T2.SlpName AS 'SalesPerson'
, T2.U_INE_Branch AS 'Branch'
, ISNULL(CAST(SUM(T0.LineTotal) AS decimal(15,2)), 0) AS 'TotalSales'
, ISNULL(CAST(SUM(T0.StockValue) AS decimal(15,2)), 0) AS 'StockValue'

FROM AU.dbo.RIN1 T0
INNER JOIN AU.dbo.ORIN T1 ON T1.DocEntry = T0.DocEntry
LEFT JOIN AU.dbo.OSLP T2 ON T2.SlpCode = T1.SlpCode

WHERE T1.DocDate >= '2014-08-01' AND T1.DocDate <= '2014-09-01'

GROUP BY T1.SlpCode, T2.SlpName, T2.U_INE_Branch

) AS T1
ON T1.SalesPerson = T0.SalesPerson

Currently I am not getting correct results (which I can do through a UNION), but am instead seeing NULLs scattered throughout, as shown below...



(I have blacked out certain details to keep the company I work for anonymous)

My question is: are NULLs unavoidable in this scenario simply because of the nature of JOINs? I have tried a 'JOIN' rather than a 'FULL JOIN' but instead of seeing all of the Sales People I only see Sales People who do not have NULLs, which looks nice but is not particularly accurate.

Any help and suggestions here will be greatly appreciated.

Kind Regards,

David
Post #1609885
Posted Wednesday, September 3, 2014 2:02 AM


SSCertifiable

SSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiable

Group: General Forum Members
Last Login: Thursday, December 18, 2014 1:43 AM
Points: 6,890, Visits: 14,254
david.dartnell (9/3/2014)
Hi Everyone,

I am in the midst of writing a report which shows the 'Total Sales', 'Gross Profit', and 'Gross Profit %' by Sales Person. Traditionally I have used a UNION between two result sets to achieve this task. However I am wondering it if is possible to achieve the same outcome with a JOIN as shown below?

...


Hi David

It looks like your suggestion will work for you, however you will have to account for null values from both tables. Try changing the outer query to this
SELECT T0.*, '#' '#', T1.*
to see what I mean. A full outer join will return matching rows between the two tables, and rows from both tables where there's no match in the other. Where there's no match, the output columns will be null.
Also, I think you should either include branch in your join, or exclude it from the GROUP BY.


“Write the query the simplest way. If through testing it becomes clear that the performance is inadequate, consider alternative query forms.” - Gail Shaw

For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
Exploring Recursive CTEs by Example Dwain Camps
Post #1609900
Posted Wednesday, September 3, 2014 6:52 PM
Valued Member

Valued MemberValued MemberValued MemberValued MemberValued MemberValued MemberValued MemberValued Member

Group: General Forum Members
Last Login: Sunday, December 14, 2014 1:08 PM
Points: 60, Visits: 126
Hi ChrisM@Work,

Thank you for your suggestions. I have now managed to find the line in my query which is returning NULL values.

(T0.TotalSales - T0.StockValue) / NULLIF(T0.StockValue, 0) AS 'Gross Profit %'

I have identified that the NULLIF(T0.StockValue, 0) section of the above line is responsible, even more specifically the T0.StockValue column itself seems to be the source of my challenge. If I modify the above line by replacing T0.StockValue with 0 then I get nothing but NULLs in my result set.

e.g.: NULLIF(0, 0) gives nothing but NULLs

On the other hand if I substitute the number 1 in place of T0.StockValue I get NO NULLs.

e.g.: NULLIF(1, 0) gives a complete set of values, with NO NULLs.

I have used a similar line of SQL in another query and it did not return any NULLs, the line is as follows -

(sum(T0.LineTotal - T0.StockValue) / nullif(sum(T0.StockValue), 0)) * 100 as 'Profit %'

Above I am summing a collection of values so I suppose there is less likely to be a NULL?

At any rate if you have any more suggestions please let me know.

Kind Regards,

David
Post #1610325
Posted Thursday, September 4, 2014 2:00 AM


SSCertifiable

SSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiableSSCertifiable

Group: General Forum Members
Last Login: Thursday, December 18, 2014 1:43 AM
Points: 6,890, Visits: 14,254
david.dartnell (9/3/2014)
Hi ChrisM@Work,

Thank you for your suggestions. I have now managed to find the line in my query which is returning NULL values.

(T0.TotalSales - T0.StockValue) / NULLIF(T0.StockValue, 0) AS 'Gross Profit %'

I have identified that the NULLIF(T0.StockValue, 0) section of the above line is responsible, even more specifically the T0.StockValue column itself seems to be the source of my challenge. If I modify the above line by replacing T0.StockValue with 0 then I get nothing but NULLs in my result set.

e.g.: NULLIF(0, 0) gives nothing but NULLs

On the other hand if I substitute the number 1 in place of T0.StockValue I get NO NULLs.

e.g.: NULLIF(1, 0) gives a complete set of values, with NO NULLs.

I have used a similar line of SQL in another query and it did not return any NULLs, the line is as follows -

(sum(T0.LineTotal - T0.StockValue) / nullif(sum(T0.StockValue), 0)) * 100 as 'Profit %'

Above I am summing a collection of values so I suppose there is less likely to be a NULL?

At any rate if you have any more suggestions please let me know.

Kind Regards,

David


Hi David

You've gone off on a bit of a tangent here so we'll deal with this first.

Function NULLIF() takes two parameters, p1 and p2. If they are different then p1 is returned, otherwise NULL is returned with data type same as p1.

Removing NULLIF from your expression above yields this:

sum(T0.LineTotal - T0.StockValue) / sum(T0.StockValue)

- which will raise a "divide by zero" error if sum(T0.StockValue) evaluates to 0. Replacing that 0 with NULL quietly yields NULL as the result of the expression. That's what the NULLIF is for - letting the expression complete without raising an error, when sum(T0.StockValue) evaluates to 0.
If you don't like NULL appearing in the result, then change it to 0 using ISNULL(), wrapped around the whole expression like this:

ISNULL((sum(T0.LineTotal - T0.StockValue) / NULLIF(sum(T0.StockValue), 0)) * 100,0)

No NULL, no error, no incorrect result.


“Write the query the simplest way. If through testing it becomes clear that the performance is inadequate, consider alternative query forms.” - Gail Shaw

For fast, accurate and documented assistance in answering your questions, please read this article.
Understanding and using APPLY, (I) and (II) Paul White
Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden
Exploring Recursive CTEs by Example Dwain Camps
Post #1610397
Posted Thursday, September 4, 2014 5:17 PM
Valued Member

Valued MemberValued MemberValued MemberValued MemberValued MemberValued MemberValued MemberValued Member

Group: General Forum Members
Last Login: Sunday, December 14, 2014 1:08 PM
Points: 60, Visits: 126
Thanks Chris, your explanation is very helpful. I now have a much better understanding of NULLIF and ISNULL.
Post #1610706
« Prev Topic | Next Topic »

Add to briefcase

Permissions Expand / Collapse