>However, under the documentation, the NOFILTER clauyse is supposed to force the physical creation on disk. As Dragan said, we dont' actually have the physical table at all in this case. Well, to be really on the safe side ... "at any time the only reliabel way to access a cursor is via its alias" ... WOn't know what happens when I port the VFP5 code to 6 with the USE DBF() AGAIN scattered all over the place.

It doesn't actually say "physical creation," it just says temp to disk. Perhaps they use the word "disk" very loosely, as in virtual memory :)

Anyway, I've been using the same USE AGAIN technique in 6, and everything is working just as in 5, very smoothly. You should have no problems with it...