But why does the table allocation 'double' when adding a clustered index?

The only explanation I can think of, is that SQL server has the information 'twice' in the database:

- ones...

It would be nice that MSSQL tables would 'be' clustered or non clustered, but in fact they 'are' heaps with an optional 'clustered index' added to then.

Removing the clustered index,...

- That's using NTFS mounting points a presume?

Of course - thanks for mentioning this (sometimes one types faster then on thinks :-P)

Why not just 'create' the G: drive?

- ex: subst g: <anypath>

Then one does not have to run the SQL in Single User mode and such....

Hmm, I probebly have some setting different, because when I only 'click' the link nothing happens....

I need to press CTRL in order te actually click the link.

So it looks like...