Traversing large record set

We're building a database to serve a web site that may have to accommodate a large number of concurrent users (up to 1000). Each user needs to traverse the results of a select statement that could return up to a million rows.

What is the most efficient way to do this? If cursors are the answer, is there a way to move backward through a result set using a cursor (e.g., "fetch prior").

We're running Oracle 8.1.7 under Solaris and Linux.

Thanks.

Paul Dalberth Received on Mon Jun 04 2001 - 18:20:04 CDT