HTH. Was fun (read a freaking pain in the @$$) to make that cumputed column indexable ... but still worth it

Could someone post the whole thing (create table, insert date and some test).. I'd like to save that somewhere for safekeeping... and actually understanding what's happening

Yup... getting tired here. Too many questions answered and not enough good sleep .

I just had some fun with that idea. If you're extremely concerned about performance then you can do something like this :

IF Object_id('Employes2') > 0

DROP TABLE Employes2

GO

IF object_id('fnEmployes2')...

This script doesn't work for me... I'm sure it's nothing but I can't find it.

Tx noeld.. didn't have the heart to try to rewrite that one...

Is it possible to try to opposite?

Like have a job that runs every 5 minutes from let's say 9 to 10, and if at 10, the files are not there,...

I feel for him too... I once lost about 3 months a data for a single very non critical column and I still have shivers about that...

At least he's only losing 12 hours of productions because he had a backup of last friday...

Sure, as long as it allows your users to do what they need.

Here's a logic that works well.

The customer table keeps the current information (no need to have that in 2 places)

On any update/delete of that information, a trigger inserts the old...

One thing that me be self evident at this point... maybe you should forbid the number 1 since 1+1+1 n times will always eventually equal any integer you want

Why do you want to do that???

Also I forgot the usual :

Table are sets, sets have no order. There's no first or last or middle in a table. If you want order specify...

Try with your original design. You'll see the problems you encounter and you'll gain experience from that.