Looking to optimize multiple JOINs into one dynamic JOIN - Possible? - sql

Question

So I currently have a query that pulls in buyer and seller names into a single column value. I am utilizing a CASE statement for if one of the buyer or seller values is NULL, and also since the user can delete the buyer or seller and it retains its original Sequence value (1, 2, 3, etc..) So for instance, say I have the following entries in the BuyerSeller table:
OrdersID | Sequence | BuyerSellerType | FormalName
------------------------------------------------------
1 | 1 | 0 | Billy Bob
1 | 2 | 0 | Sally Sue
1 | 1 | 1 | Joe Dirt
1 | 3 | 1 | Dwayne Johnson
My SELECT statement below will return the following:
BuyerFormalName | SellerFormalName
-----------------------------------------------
Billy Bob, Sally Sue | Joe Dirt, Dwayne Johnson
Here's the statement. Now, I am just trying to see if there is any way to dynamically or more easily produce the same results without the need for so many JOINs? I am getting the results I want, but I am looking to fine tune my skills and knowledge, especially since I might have to potentially add >3 more JOINs. Any insight would be very much appreciated!
SELECT
CASE
WHEN ISNULL(B.FormalName,'') <> '' AND ISNULL(B2.FormalName,'') <> '' AND ISNULL(B3.FormalName,'') <> ''
THEN B.FormalName + ', ' + B2.FormalName + ', ' + B3.FormalName
WHEN ISNULL(B.FormalName,'') <> '' AND ISNULL(B2.FormalName,'') = '' AND ISNULL(B3.FormalName,'') <> ''
THEN B.FormalName + ', ' + B3.FormalName
WHEN ISNULL(B.FormalName,'') <> '' AND ISNULL(B2.FormalName,'') <> '' AND ISNULL(B3.FormalName,'') = ''
THEN B.FormalName + ', ' + B2.FormalName
WHEN ISNULL(B.FormalName,'') = '' AND ISNULL(B2.FormalName,'') <> '' AND ISNULL(B3.FormalName,'') <> ''
THEN B2.FormalName + ', ' + B3.FormalName
WHEN ISNULL(B.FormalName,'') = '' AND ISNULL(B2.FormalName,'') <> '' AND ISNULL(B3.FormalName,'') = ''
THEN B2.FormalName
WHEN ISNULL(B.FormalName,'') = '' AND ISNULL(B2.FormalName,'') = '' AND ISNULL(B3.FormalName,'') <> ''
THEN B3.FormalName
ELSE B.FormalName
END AS 'BuyerFormalName'
,CASE
WHEN ISNULL(S.FormalName,'') <> '' AND ISNULL(S2.FormalName,'') <> '' AND ISNULL(S3.FormalName,'') <> ''
THEN S.FormalName + ', ' + S2.FormalName + ', ' + S3.FormalName
WHEN ISNULL(S.FormalName,'') <> '' AND ISNULL(S2.FormalName,'') = '' AND ISNULL(S3.FormalName,'') <> ''
THEN S.FormalName + ', ' + S3.FormalName
WHEN ISNULL(S.FormalName,'') <> '' AND ISNULL(S2.FormalName,'') <> '' AND ISNULL(S3.FormalName,'') = ''
THEN S.FormalName + ', ' + S2.FormalName
WHEN ISNULL(S.FormalName,'') = '' AND ISNULL(S2.FormalName,'') <> '' AND ISNULL(S3.FormalName,'') <> ''
THEN S2.FormalName + ', ' + S3.FormalName
WHEN ISNULL(S.FormalName,'') = '' AND ISNULL(S2.FormalName,'') <> '' AND ISNULL(S3.FormalName,'') = ''
THEN S2.FormalName
WHEN ISNULL(S.FormalName,'') = '' AND ISNULL(S2.FormalName,'') = '' AND ISNULL(S3.FormalName,'') <> ''
THEN S3.FormalName
ELSE S.FormalName
END AS 'SellerFormalName'
FROM
Checks C
LEFT JOIN BuyerSeller B
ON C.OrdersID = B.OrdersID
AND B.Sequence = 1
AND B.BuyerSellerType = 0
LEFT JOIN BuyerSeller B2
ON C.OrdersID = B2.OrdersID
AND B2.Sequence = 2
AND B2.BuyerSellerType = 0
LEFT JOIN BuyerSeller B3
ON C.OrdersID = B3.OrdersID
AND B3.Sequence = 3
AND B3.BuyerSellerType = 0
LEFT JOIN BuyerSeller S
ON C.OrdersID = S.OrdersID
AND S.Sequence = 1
AND S.BuyerSellerType = 1
LEFT JOIN BuyerSeller S2
ON C.OrdersID = S2.OrdersID
AND S2.Sequence = 2
AND S2.BuyerSellerType = 1
LEFT JOIN BuyerSeller S3
ON C.OrdersID = S3.OrdersID
AND S3.Sequence = 3
AND S3.BuyerSellerType = 1

Answer 1

Are you looking for something like this?
WITH BuyerSeller
AS
(
SELECT *
FROM
(
VALUES (1,1,0,'Billy Bob'),
(1,2,0,'Sally Sue'),
(1,1,1,'Joe Dirt'),
(1,3,1,'Dwayne Johnson')
) A(OrdersID,[Sequence],BuyerSellerType,FormalName)
)
SELECT DISTINCT OrdersID,
STUFF(Bname,1,2,'') AS BuyerFormalName,
STUFF(Sname,1,2,'') AS SellerFormalName
FROM BuyerSeller A
CROSS APPLY (
SELECT ', ' + FormalName
FROM BuyerSeller B
WHERE A.OrdersID = B.OrdersID
AND BuyerSellerType = 0
ORDER BY [Sequence]
FOR XML PATH('')
) CA(Bname)
CROSS APPLY (
SELECT ', ' + FormalName
FROM BuyerSeller B
WHERE A.OrdersID = B.OrdersID
AND BuyerSellerType = 1
ORDER BY [Sequence]
FOR XML PATH('')
) CA2(Sname)
Results:
OrdersID BuyerFormalName SellerFormalName
----------- --------------------- -----------------------
1 Billy Bob, Sally Sue Joe Dirt, Dwayne Johnson

Answer 2

I still prefer old-style MAX/CASE in such a case if the number of rows/values is known (and small):
SELECT
c.OrdersID,
BuyerFormalName,
SellerFormalName
FROM Checks C
LEFT JOIN
(
SELECT
OrdersID,
STUFF(MAX(CASE WHEN Sequence = 1 AND BuyerSellerType = 0 THEN ', ' + FormalName ELSE '' end) +
MAX(CASE WHEN Sequence = 2 AND BuyerSellerType = 0 THEN ', ' + FormalName ELSE '' end) +
MAX(CASE WHEN Sequence = 3 AND BuyerSellerType = 0 THEN ', ' + FormalName ELSE '' end), 1,2,'') AS BuyerFormalName,
STUFF(MAX(CASE WHEN Sequence = 1 AND BuyerSellerType = 1 THEN ', ' + FormalName ELSE '' end) +
MAX(CASE WHEN Sequence = 2 AND BuyerSellerType = 1 THEN ', ' + FormalName ELSE '' end) +
MAX(CASE WHEN Sequence = 3 AND BuyerSellerType = 1 THEN ', ' + FormalName ELSE '' end), 1,2,'') AS SellerFormalName
FROM BuyerSeller
GROUP BY OrdersID
) AS b
ON C.OrdersID = B.OrdersID
;