>We have a situation where we are using ORACLE as a back-end to our VFP middle-tier servers.
>
>Our source data is being updated every 3 seconds or so (could be 20 - 30 inserts, updates, changes, deletions), and we plan to have VFP remote views scheduled to run at particular time intervals.
>
>Since the source (ORACLE) and middle-tier are 4500 miles apart(!) and our source data runs into millions of records, we do not want to rebuild the remote views with every refresh. (Our networks guys would have our guts for garters!)Rather, we will either REFRESH() or REQUERY() the views. Could someone give me an insider's/hacker's look into those two functions please?

Distance is practically meaningless over TCP/IP.

>I'm particularly interested in the data sets each of them return, and which one would tax networks the least (in terms of whether just updates are sent through, or the whole view itself etc.)
>

Refresh does nothing with respect to data retrieval. Requery will re-submit the SQL to the server.

>Also, would I be better off if I used SQL Passthrough? ODBC or ADO? I have not yet decided. Any comments, experiences will be appreciated.
>
>Thanks very much in advance

I use remote views for updateable data. I use SPT to create cursors for read-only data. Both these methods use ODBC where I use the MS ODBC for Oracle exclusively. I have not had an opportunity to use ADO yet.