>In general, int PKs perform faster than GUID. You do know that you can utilize the negative range and also if you expect your table to grow really large, you can use BigInit.
>
Yes, I'm aware of the negative part
The question about faster is how much faster
Also. I want to be able to extract a subset of the db to put on a laptop, work on it and merge/sync it back later - hence the guid
Gregory