VBScript and SQL Server 2005 compare current row to previous row - sql-server

Question

I am dealing with some legacy tech at my new job. Everything is either VBS or VBA and SQL Server 2005 is still running for at least 1 server, and 2008 on another. I can't change the tech I have to work with, so please don't respond with "Just use PowerShell" ... which is all I got from the Microsoft forums.
I need to modify an existing VBScript to iterate through the records of a table and give each page of a document a page number. The page number starts at 1 and resets to 1 every time the location folder changes, as defined by the row's value in the column Path. So I need to compare the current row's Path to the previous row's Path to see if there is a change.
From the specs request:
Use the following structure:
Outermost folder = Box field
Sub-folder = File Folder field
Docs within will each begin at "0000001"
Sub-folder = File Folder field
Docs within will each begin at "0000001"
Sub-folder = File Folder field
Docs within will each begin at "0000001"
Here is a very simple but effective Excel script that works, unfortunately I have to use VBScript so a third-party software can run the script:
=IF "Folder field in this row" = "Folder field in row above"
True = "above column" (beg doc of previous doc) + "pgcount of previous doc"
False = "1"
Here is the existing SQL statement from within the VBScript (Sorry for the naming convention - it is not my choice):
UPDATE tblpage
SET tblpage.UID = newtable.keyid
FROM tblpage inner join (SELECT PKEY, '" & prefix & "' + REPLICATE(0, " & padnum & " -
LEN(convert(varchar(" & padnum & "),(" & startnum & " + rank() OVER (ORDER
BY TBLDOC." & sortfield & ") + Page - 2)))) + CONVERT(nvarchar(" & padnum &
"), convert(varchar(" & padnum & "),(" & startnum & " + rank()
OVER (ORDER BY TBLDOC." & sortfield & ") + Page - 2))) AS keyid
FROM tblpage
inner join tbldoc on tblpage.id = tbldoc.id) as newtable ON newtable.pkey = tblpage.pkey;
variables taken from user input in VBScript - explanation
prefix - optional text prefix to the iterative number ie "SET" for "SET0000001"
padnum - number of 0s before iteration begins ie "000000" for "SET0000001"
startnum - starting number for the iteration ie "1" in "SET0000001"
sortfield - which table column to sort by ie a table's UID.
The above SQL does the iteration correctly but doesn't reset on new folder. I have been looking into SQL SERVER – How to Access the Previous Row and Next Row value in SELECT statement? – Part 2, but I am still quite confused.
My thoughts are that I join just the Path to the newtable based on the row's UID - 1, and then a case statement to compare them. If different, reset the page number. If the same, continue as normal.
Thanks!!

Answer 1

This answer comes from my old CS college roommate, but he doesn't have a Stack account(!?) so I'll post for him...
Even though the online docs for the OVER function are stamped 2008+, it actually does work just fine in 2005, as evidenced by the original posted query. This also means I can bypass all the nastiness of comparing previous row to current row by using PARTITION BY to reset on the chosen column Path
UPDATE tblpage
SET tblpage.UID = newtable.keyid
FROM tblpage inner join (SELECT PKEY, '" & prefix & "' + REPLICATE(0, " & padnum & " -
LEN(convert(varchar(" & padnum & "),(" & startnum & " + rank() OVER (
PARTITION BY TBLDOC.Path ORDER
BY TBLDOC." & sortfield & ") + Page - 2)))) + CONVERT(nvarchar(" & padnum &
"), convert(varchar(" & padnum & "),(" & startnum & " + rank()
OVER (PARTITION BY TBLDOC.Path ORDER BY TBLDOC." & sortfield & ") + Page - 2))) AS keyid
FROM tblpage
inner join tbldoc on tblpage.id = tbldoc.id) as newtable ON newtable.pkey = tblpage.pkey;