>>No it's not. IIRC, I posted improved version in UT magazine a while ago but the link http://www.utmag.com/ doesn't work.
>
>Thanks, until I reach a big number of records in that table, I will be able to use the NEWID() approach. Then, I will rely that random generator based on something else.

Check message #777074

Btw, the link still doesn't work.