Monthly Archives: May 2010

Summit 2010 Abstract Submission

  By Michelle Ufford | May 28, 2010 - 12:02 pm | , ,
1 Comment

Jeremiah Peschka just tweeted about the looming deadline to submit an abstract for Summit 2010. I’ve been trying to think of a good topic to present on, and this finally got my butt in gear to submit one! For those interested, here’s what I submitted:

Heaps of Trouble, Clusters of Glory – A Look At Index Internals

Indexes are a crucial component of SQL Server, especially in performance tuning, yet many DBA’s don’t fully understand how indexes work. In this in-depth session, Michelle will examine the anatomy of indexes, from how they’re stored to how they get fragmented. And to make our inner geeks happy, she’ll look at page data to show you what’s happening behind the scenes. Topics covered will include index structure, fragmentation and defragmentation, index partitioning, and index filtering. If you’ve ever wondered just *what* goes on in an index, don’t miss this session!

Special thanks to Chris for help with the session title. :)

DELETE 5_Useless_Things FROM [SQL Server]

  By Michelle Ufford | May 14, 2010 - 7:34 am | ,
18 Comments

It’s been a while since I’ve been caught up in a round of chainblogging, the blogosphere’s version of a Facebook meme. This time, Denis Gobo tagged me in a post started by Paul Randal. Paul asked us to list the “top-5 things in SQL Server we all wish would just be removed from the product once and for all.” I reviewed other posts, and the good and bad news is that they already listed several of the same things I would have. The good news is I’m apparently not alone; the bad news is that means I need to come up with something original! So while these wouldn’t necessarily be the *first* 5 on my list, they’d still be on the list nevertheless:

Default Autogrowth Options
Okay, so I lied. I’m not completely original. Yes, I know Paul Randal also commented on this one. While I said I would try to come up with only original ones, this one just has to be repeated. I’ve actually this option overlooked in production environments, resulting in thousands of VLF’s. It’s just a terrible default, and it needs to be changed.

Edit Top 200 Rows
This “feature” is just asking for trouble. Any DBA who is managing a SQL Server database should understand how to actually write insert/update/delete statements. Maybe leave the option available in SQL Express, but please remove it from SQL Server Standard & Enterprise.

Debug
There’s nothing wrong with the Debug option, but I think it should be removed as a default option for the toolbar. It’s easily mistaken for “Execute,” which I’ve seen more than one DBA do on occasion.

PIVOT
I understand the need to pivot your data, but let’s face it. PIVOT is a clunky, expensive SQL operation. Let’s move the presentation tasks to the presentation layer (.NET), and reserve the database layer for what it does best.
Update: By popular demand, I have removed PIVOT from this list. Who am I to argue with such fine folks? :)

Cursors
Okay, okay, I know I can’t actually get rid of this, BUT I think it gets abused way too much. Set-based operations, anyone?

Alrighty, now it’s my turn to tag! I’m not sure if they’ve already been hit, but I’m tagging: