Cetin,

SNIP
>I expect seek be fastest here (I would do 'set near on' and seek 1700002, 1700001 might not be there).
>OTOH removing 'nofilter' on SQLs I expect them to be much faster than 30 secs too.
>VFP SQL unlike SQL server's is not kind enough to break when it finds top n :( If this wasn't a PK issue it would count too.
>Cetin

Why would you think that rmoving NOFILTER would make it dramatically faster?

Also, are you saying that you believe that TOP n works substantially differently between SQL Server and VFP SQL? This would seem odd to me.

Jim Nelson