24.25 Hours of PASS
The 24 Hours of PASS kicked off at 19:45 EST yesterday and has so far, with about three hours to go, been a great technical and educational success. Presenters have been solid for the most part, and there have been limited technical issues to work out early on. PASS’s own Tom LaRock has also been hanging on through each session live streaming out on U-Stream as well.
What have the highlights been for me?
Learning more about the tools I love. It never fails that I learn something new each day and the last 24 hours have provided a wellspring.
Sessions on building better blogs (Steve Jones), the simple recovery model (Kalen Delaney), effective indexing (Gail Shaw), query tuning (Grant Fritchey), and database compatibility settings (Don Vilen) were standouts in my book. I plan on placing many ideas from these sessions into operation immediately.
For example I did not realize that when using a multiple key index, any columns listed after the first inequality column is hit can not be used in a seek operation. It was also noted by Gail Shaw that you should place your most narrow equality key first in the compound-key index, relegating less-distinct and all inequality keys afterwards.
Furthermore, Don Vilen enlightened us with what exactly compatibility-modes mean and don’t mean in SQL Server. Leaving a compatibility mode of a database set to an older version of SQL Server does NOT… NOT mean that the database remains in that version. I highly recommend you review his session when made available offline around September 9th, 2009.
That leads to the final tidbit of information we’ve been fed. Select (stress that SELECT) sessions are going to be made available from recordings as soon as that September 9th date I mentioned. Stay tuned to the PASS website for details on which session will be available and when they are available.
Now it’s time to get back for the last 2 sessions!
It has been a very great event. I’ve managed to watch a lot of the sessions and learn a ton of new things.
I hope they think about maybe doing smaller version of this in the future.
Oops, looks like I wasn’t clear in what I was saying. Sorry.
Columns after the first inequality aren’t ignored, they just can’t be used for a seek operation. Order of columns in index should be based on how they’re used and on the selectivity of the column. Main thing to remember is that seeks can only be done on left-based subsets of the index key.
Fixed the post Gail. Thanks.