>>

Create cursor .....
>>Index on ... tag ...

will create an index in memory!
>
>Will it really? I am not sure about this, but I thought that perhaps VFP might decide from case to case (depending on size?) whether it is more appropriate to keep it in memory or on disk, just as it is supposed to do with the cursors themselves - or perhaps always put it on disk, with a temporary filename.

Actually, I think you'll find that VFP virtually ALWAYS creates tables/cursors/indexes in memory, only physically writihg them when it "has to".

I know that I've written multiple 50,000+ record tables with records to each table one after the other and yet the tables (new in the program) ended up each being UNfragmented!
This indicated to me that they were physically written after I was finished with them.
Of course system RAM and other RAM usage in both VFP and other active applications would affect this, but I think it's a safe bet that smaller cursors particularly will not end up on HD generally, and I would assume the same for their CDXs. NOT, however, for any related .FPTs, which are always written as records are created.