Home/ Questions/Q 6535039
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-25T10:19:27+00:00

This is SQL Server 2000 so I don’t have any windowing functions (row_number). I

  • 0

This is SQL Server 2000 so I don’t have any windowing functions (row_number).

I have a table emp_data :

emp_id     datime                 miles    gallons
23148      2011-08-21 02:00        32       3
23148      2011-08-21 09:00        38       4
23148      2011-08-21 11:00        40       5
42938      2011-08-20 03:00        23       1
42938      2011-08-22 08:00        53       13
  • Each row is cumulative (running?) from the previous one.

I need to get the number of miles driven by the employee, which I do by subtracting the miles for the earliest date minus the miles for the latest date. (40-32 = 8 miles driven for empid=23148). I need to do this for gallons too.

I need to calculate miles per gallon for each driver.

The end result should be this:

  emp_id    miles   gallons
   23148      8       2
   42938      30      12

Doing it for multiple drivers is where I’m stuck. In SQL Server 2005, I could probably do row_number partition_by, but don’t know what to do in SQL Server 2000. I’ve done something like this for one driver. Won’t work partitioned by drivers. Had to use identity() in place of row_number.

SELECT IDENTITY(int) as id, emp_id, datime, miles, gallons
into #t1
FROM emp_data
where 
   emp_id='18018'
   and datime >= '20110820 02:00'
   and datime <= '20110827 02:00'
ORDER BY datime

select foo1.emp_id,foo2.miles - foo1.miles as miles_driven,
    foo2.gallons - foo2.gallons as gallons_used
from (
    SELECT * 
    FROM #t1
    where id = 1) foo1
CROSS JOIN  (
    SELECT *
    from #t1 
    where id = (select max(id) from #t1 t)
    ) foo2

I do have a linked server to the SQL Server 2000 db from SQL Server 2008 so I’m thinking of getting the data and then processing there, but there’s about 1 million records for just one week. I might need to do this for YTD.

Let me know if something is unclear. Sorry I don’t have any sample data.

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    I think this is what you’re looking for (no need for temp tables) but only use if you have additional data constraints that we’re not seeing. Otherwise go with smdrager answer

    SELECT minmaxdate.emp_id, 
       LAST.miles - FIRST.miles, 
       LAST.gallons - FIRST.gallons 
FROM   (SELECT emp_id, 
               MIN(datime) firstdate, 
               MAX(datime) lastdate 
        FROM   emp_data 
        GROUP  BY emp_id) minmaxdate 
       INNER JOIN emp_data FIRST 
         ON FIRST.emp_id = minmaxdate.emp_id 
            AND FIRST.datime = minmaxdate.firstdate 
       INNER JOIN emp_data LAST 
         ON FIRST.emp_id = minmaxdate.emp_id 
            AND LAST.datime = minmaxdate.lastdate
    • 0