Todd,

I'd personally consider this to be an "issue" with VFP's internal choice of using a "filtered result set" and it has long been my opinion that VFP should NEVER do that!!!

I tried it out on a table, much as Garrett did in another message (though I took my cue from YuriR).
When I did a SELECT * that didn't involve Rushmore then I got records that were Deleted() in the original but they showed as UNdeleted in the result set (which I believe is the 'normal' expectation).
Doing the same with a Rushmore optimizeable WHERE did give me Deleted records that showed as DELETED because what I really got was a filtered result set. Adding the NOFILTER clause to this last one produced records (including deleted in original ones) that showed as UNdeleted.

In summary, I believe that COUNTING on getting the DELETED() flag in any Select-SQL is a bad thing to do.

cheers

>>Does the user also need to be able to recall a previously deleted record?
>
>Yes.
>
>Seems odd that, as widely promoted as updatable views are, that at bug this obvious would still be around. I cant help thinking there must be something Im missing. Think Ill check the MSDN Knowledge Base.