Limiting selected rows in SQL Server - sql-server-2008

Question

In the query below I want to retrieve #MaximumRecords rows, so that no ProjectId will have rows left out beyond #MaximumRecords.
For example if #MaximumRecords=100, and ProjectId=7 has records at rows number 99-102, I wish to retrieve only rows with ProjectId=1 to ProjectId=6 (The query will run again later starting at ProjectId=7). How do I do that?
SELECT TOP (#MaximumRecords)
t1.ProjectId,
t1.Row2,
t2.Row3
FROM Table1 t1
JOIN Table2 t2 ON t1.ProjectId = t2.ProjectId
ORDER BY
t1.ProjectId ASC
WHERE
t1.ProjectId > #InitialProjectId

Answer 1

I worked up a solution using the Sales.SalesOrderHeader and Sales.SalesOrderDetail tables in the AdventureWorks2008R2 database using the technique to get a running total described here.
The basic idea is to get the running total of the count for each SalesOrderID (ProjectID in your case) and then select all of the data for each SalesOrderID where the running total of the count is less than #MaximumRecords. You would then need to capture the maximum ID in the data returned and use that value in the next run of your query.
This task gets a little easier with SQL Server 2012 which is also described in the link given above.
Here it is...
USE AdventureWorks2008R2
IF OBJECT_ID('tempdb..#Test', 'U') IS NOT NULL DROP TABLE #Test;
DECLARE #MaximumRecords INT
DECLARE #InitialSalesOrderID INT
SET #MaximumRecords = 500
SET #InitialSalesOrderID = 43663
SELECT a.SalesOrderID, COUNT(*) AS 'Count'
INTO #Test
FROM Sales.SalesOrderHeader a
INNER JOIN Sales.SalesOrderDetail b ON a.SalesOrderID = b.SalesOrderID
WHERE a.SalesOrderID > #InitialSalesOrderID
GROUP BY a.SalesOrderID
SELECT * FROM Sales.SalesOrderHeader a
INNER JOIN Sales.SalesOrderDetail b ON a.SalesOrderID = b.SalesOrderID
WHERE a.SalesOrderID IN (
SELECT
a.SalesOrderID
FROM
#Test a
WHERE (
SELECT
SUM(Count)
FROM
#Test b
WHERE
b.SalesOrderID <= a.SalesOrderID
) < #MaximumRecords
)
Noel