OK Marcel,IC.

I certainly understand your frustration, and have run into this often enough myself <s>. (For the same reason as you, usually, SELECT en route to another data format <s>.)

But I'm kinda glad that the cursoradapter class isn't a "magic bullet" for it.

For one thing, it might mean that cursoradapter wasn't using "real" VFP SELECT-SQL under the covers. That would be bad, because it would mean that there might be a disconnect if a new feature were added to SELECT, etc, or even if no new feature, perhaps not all the clauses and current special-to-VFP features would work (UDFs come to mind!).

For another... if this behavior were ever changed, I'd want it changed for SELECT-SQL, wouldn't you? Not just for cursoradapter <s>.

Later,

>L<