>>>Kevin,
>>>
>>>>
>>>>Let's compare:
>>>>
>>>>APPEND FROM
>>>>
>>>>Records/Time
>>>>
>>>>17/0.040
>>>>40/0.010
>>>>15/0.010
>>>>22/0.010
>>>>16/0.010
>>>>22/0.010
>>>>31/0.010
>>>>
>>>>SCATTER/INSERT
>>>>
>>>>17/0.040
>>>>40/0.010
>>>>15/0.020
>>>>22/0.010
>>>>16/0.020
>>>>22/0.010
>>>>31/0.010
>>>>
>>>>Do you tink that's massive difference?
>>>>
>>>
>>>Am I reading this right ? Number on left represent record counts ? If yes then testing with so few low reccounts and total time taking 14secs ??? Well compared to (was done in 1990's) from multiple 80Mb+ tables in fox2x doing 'contains' searches at most in 30 secs on a Novell network with 1Mb RAM 80286 boxes that's really slow. Do not use VFP.
>>
>>No, the 14 seconds came from quering 7 "remote" locations, these tests are for 7 local dbs that are being queried
>
>
>Still unclear. Local or remote do tables have 17-31 records per table ?

No, that is the result of the query, so the scatter/append from had to perform the operation on 17 records.

>
>>
>>>>
>>>>That will surely depend on the design of the app, maybe if a full blown OO design was adopted there wouldn't be so much need for data-intensive tasks, if there was no way around it then the back-end SP's could handle it.
>>>>
>>>>Kev
>>>
>>>Well I thought I made a OO design but who knows.
>>
>>That's just my opinion, I've found that if a full-blown oo approach is taken then the app becomes less data-intensive...
>>
>>Kev
>
>Yes that's your opinion.
>Cetin