• Which is the best way to perform with tables to do DML ?

    As with so many things in SQL Server, the answer is "it depends". But as a general rule, UNIQUEIDENTIFIERS are typically a poor choice for a primary key that is also defined as a clustered index for a couple of reasons.

    1. The key is included as the lookup value for all non-clustered indexes and a GUID is significantly wider than an Int or a BigInt. That will require more disk space to accomodate the increased data type size.

    2. UNIQUEIDENTIFIER's as the clustered index are prone to fragmentation because they are not sequential and are generated in a random order.

    You can read more on Kimberly Tripp's blog here: http://www.sqlskills.com/blogs/kimberly/post/GUIDs-as-PRIMARY-KEYs-andor-the-clustering-key.aspx

    If you're going to use a UNIQUEIDENTIFIER as a primary key / clustered index, you want to make sure you are using the NEWSEQUENTIALID() function.