Posted by Alex Kuznetsov on 07/16/07 13:47

On Jul 16, 4:00 am, alexander.arvids...@gmail.com wrote:
> Hello,
>
> I'm new to this group and I sincerely hope I'm not stepping on anyones
> toes or doing something the wrong way around by beginning my time here
> by asking a question.
>
> I'm a Oracle DBA from the beginning (been one since '97) and I've been
> using SQL Server since 2001. Yesterday one of my customers (I'm a
> consultant) showed me a problem they have, and it turns out it is the
> 'Sparse Extent Scenario' (seehttp://sqlforums.windowsitpro.com/web/forum/messageview.aspx?catid=25...
> and scroll down to the user cmt_SQL)
>
> The solution is hence simple; add clustered indexes to those tables
> that don't have them. But, here is the actual problem:
>
> The creators of the software that my customer uses (two different
> systems) BOTH claim that using clustered indexes hampers performance,
> each and every time. I can't find ANY resource on the internet that
> validates this, quite the opposite. I am told that the best practices
> is to always us a clustered index on a table.
> Following their own guidelines, there is no clustered index in sight,
> and hence some tables have a whopping 30GB(!) of unused space.
>
> I'm looking for ammunition to use on the abovementioned developers.
> I'm looking for detailed technical explanations why a clustered index
> is so much better than an unclustered ditto. I suspect I would find it
> in Kalen Delaney's books, but unfortunately I don't have them before
> me (although I'm looking to order them). Could anyone point me to a
> suitable usenet post, a web page or anything similar?
>
> Kind regards,
> Alexander

In some that's correct. Refer to a series of excelletn articles by
Greg Linwood, SQL Server MVP:
http://blogs.sqlserver.org.au/blogs/Greg_Linwood/

Note that you can create a clustered index and drop it immediately -
that will take care of your fragmentation.

Alex Kuznetsov, SQL Server MVP
http://sqlserver-tips.blogspot.com/

[Back to original message]