Question

The problem is that we are to get total number of rows before and after a given row (identified by, say, primary key).

I have tried following in T-SQL (MSSQL 2008). It is giving right result, but I don’t know if this is best way to do it.

;WITH cte_before AS 
(
    SELECT ROW_NUMBER() OVER ( Order By CustomerId ) [Row Number], customerid, 
           firstName 
    FROM SalesLT.Customer
 ),
cte_nums AS 
(
    SELECT ROW_NUMBER() OVER ( Order By CustomerId ) [Row Number1] 
    FROM SalesLT.Customer
)
SELECT [Row Number]-1 [before], MAX([Row Number1]) - [Row Number]
       , CustomerID, FirstName 
FROM cte_nums, cte_before
GROUP BY [Row Number], CustomerID, FirstName
HAVING CustomerID = 55

How can we improve it in T-SQL and how can we accomplish in other SQL dialects and server (like Oracle, MySQL, sqlite, FireBird etc.)

Answer 1

I think this will work in any dialect:

select 
   (select count(*) from SalesLT.Customer where customerid < 55) as rows_before,
   (select count(*) from SalesLT.Customer where customerid > 55) as rows_after,
   CustomerID, FirstName
from SalesLT.Customer
where CustomerID = 55