Declare, Set, and Use Variable in SQL Query (not a stored proc) - sql

Question

Maybe I'm missing something simple, but is there a way to declare a variable and use it within a single SQL query? That way, if you have a piece of information that is generated using a large nested query, you don't have to keep running that same query over and over again.
This would be a decent example illustrating the problem:
SELECT id, CASE WHEN LEN(HUGESUBQUERY) > 10 THEN
LEFT(HUGESUBQUERY, LEN(HUGESUBQUERY)-2) ELSE HUGESUBQUERY END AS result FROM table
With everything added in, it gets really long. Here is the actual query I am working with. (of note, I am running this from Dynamics AX, so I have no flexibility on doing anything other than everything in one SELECT statement. None of the solutions with DECLARE, SET, etc. are available in this environment.)
SELECT (CAST (
CASE WHEN
RIGHT(
(CONCAT(COALESCE((SELECT TOP 1 SalesId FROM CustInvoiceJour WHERE
CustInvoiceJour.InvoiceId = T1.INVOICE), 'NONE'),
'-', ROW_NUMBER () OVER (PARTITION BY T1.ACCOUNTNUM, (SELECT TOP 1 SalesId
FROM CustInvoiceJour
WHERE CustInvoiceJour.InvoiceId = T1.INVOICE) ORDER BY T1.DUEDATE)-1))
, 2) = '-0'
THEN
LEFT(
(CONCAT(COALESCE((SELECT TOP 1 SalesId FROM CustInvoiceJour WHERE
CustInvoiceJour.InvoiceId = T1.INVOICE), 'NONE'),
'-', ROW_NUMBER () OVER (PARTITION BY T1.ACCOUNTNUM, (SELECT TOP 1 SalesId
FROM CustInvoiceJour
WHERE CustInvoiceJour.InvoiceId = T1.INVOICE) ORDER BY T1.DUEDATE)-1)),
LEN(
(CONCAT(COALESCE((SELECT TOP 1 SalesId FROM CustInvoiceJour WHERE
CustInvoiceJour.InvoiceId = T1.INVOICE), 'NONE'),
'-', ROW_NUMBER () OVER (PARTITION BY T1.ACCOUNTNUM, (SELECT TOP 1 SalesId
FROM CustInvoiceJour
WHERE CustInvoiceJour.InvoiceId = T1.INVOICE) ORDER BY T1.DUEDATE)-1))
)-2
)
ELSE
(CONCAT(COALESCE((SELECT TOP 1 SalesId FROM CustInvoiceJour WHERE
CustInvoiceJour.InvoiceId = T1.INVOICE), 'NONE'),
'-', ROW_NUMBER () OVER (PARTITION BY T1.ACCOUNTNUM, (SELECT TOP 1 SalesId
FROM CustInvoiceJour
WHERE CustInvoiceJour.InvoiceId = T1.INVOICE) ORDER BY T1.DUEDATE)-1))
END
AS NVARCHAR(21))) AS DOC_ID,
T1.ACCOUNTNUM AS CUSTOMER_ID,T1.AMOUNTCUR AS DOCTOTAL,T1.DUEDATE AS
DUEDATE,T1.DOCUMENTDATE AS DOCDATE,
T1.RECID AS ID,T1.DATAAREAID AS DATAAREAID,T1.PARTITION AS
PARTITION,T1.RECID AS RECID,
(CAST ((T1.RECID + T1.RECVERSION) AS BIGINT)) AS SWX_COMPUTEDVERSION FROM
CUSTTRANS T1 WHERE (AMOUNTCUR>0)
ORDER BY DOC_ID

Answer 1

Yes you can considering that the query is not Correlated. Try this
DECLARE #HUGESUBQUERY VARCHAR(500)
SET #HUGESUBQUERY = (SELECT TOP 1 something
FROM somewhere
INNER JOIN table2
ON table2.id = somwhere.id
INNER JOIN table3
ON table3.id = table2.id
WHERE condition1 = condition1
AND condition2 = condition2)
SELECT id,
CASE #HUGESUBQUERY
WHEN Len(#HUGESUBQUERY) > 10 THEN LEFT(#HUGESUBQUERY,
Len(#HUGESUBQUERY) - 2)
END AS result
FROM table

Answer 2

For a correlated subquery, you can use CROSS APPLY. Since we don't have a real query, I'll just put one that demonstrates the technique:
select
*,cname
from
sys.objects so
cross apply
(select top 1 sc.name from sys.columns sc where sc.object_id = so.object_id) t (cname)
And now cname can be used throughout the SELECT clause to refer to the name value returned from sys.columns.