>This afternoon, and thanks to those who helped me with those SQL, I was able to define the required SQL I was looking for. This works. Basically, the actual command is:
>
>

>SELECT RO.FirstName
>FROM (SELECT TOP 10 Event.RO_No,Min(Event.DateTime) AS LastUpdate
> FROM Event
> INNER JOIN RO ON Temp.RO_No=RO.RO_No
> GROUP BY Event.RO_No
> ORDER BY MIN(Event.Datetime)) Temp
>INNER JOIN RO ON Temp.RO_No=RO.RO_No
>

>
>So, this returns something like this:
>
>

>Ro_No       DateTime
>  1         2007-12-17 14:41:25.704
>  3         2007-12-17 14:42:25.704
>  7         2007-12-17 14:43:25.704
> 22         2007-12-18 15:41:25.704
> 23         2007-12-19 16:41:25.704
> 24         2008-12-20 07:41:25.704
> 25         2009-12-17 08:41:25.704
> 46         2009-12-17 05:41:25.704
> 48         2009-12-18 05:41:25.704
> 49         2009-12-19 14:41:25.704
>

>
>But, Ro_No, in the Event table appears in multiple records. The reason is that the Event table logs everything when there is a change in the record. So, I have a robot process which processes 10 records at a time. So, in the next loop, it will execute the same SQL again and obtain a new batch of 10 records. However, in that new batch, I may end up with a RO_No that I already processed. In our design, there is only a need to process the last one from the Event table. I do not know if that is possible here, if that would require a derived SQL or else.
>
>The problem I have presently is that I have about 2000 records to process. So, this would make it 200 times. But, I only have 56 records in the RO table. So, at most, if I can refine this SQL, I would only have to do 6 loops.

Why not to add a bit field IsProcessed (or Processed) to RO table? When you do the work in your robot process you update the flag and therefore you will not select same records again.