>>After I have the data from the Oracle DB in a CA cursor, is the sys(3054) function usable for subsequent queries?
>
>Yes. As stated before it only works with Native DBF (tables or cursors). If you query data from Oracle and now you have it locally (dbf) it will then report on it. Of course you'd have to index that local data to gain any advantages from Rushmore.
>
>As generally queried data is supossed to be a fairly small set (you would not bring tens of thousands of records from Oracle as that may not make sense), the gains you may have from Rushmore-optimizing your resulting cursor may be minimal. YMMV.

The gains may actually be major losses.

http://msvfp.advisorguide.com/doc/01468

Basically LOCATE and SQL against a small data set will work quickly enough. Adding the indexes first will eat up time. Not adding the indexes makes the Rushmore attempt wasteful. So SET OPTIMIZE off or LOCATE NOOPTIMIZE to get the best performance.

I add the indexes to resulting cursors so the user can reorder data displayed in a grid without reselecting the data and ordering it via the query. Since they are also rarely the same indexes I'd need for Rusmore I continue to use the NOOPTIMZE clause.