Thomas-282729 (1/1/2011)

---

Inefficient with your data types is a relative problem. Is storing the year out to its full four digits inefficient? It was 40 years ago. You are discounting development...

Steve Jones - SSC Editor (1/1/2011)

---

CPU power has increased, but a great talk from Dr. David Dewitt in 2009 at the PASS Summit showed that disk xfer rates have not....

GPO (1/1/2011)

---

What my argument doesn't address here is the opportunity cost of parsimony. Does it cost more to hire competent DBAs than it does to get the same outcome through...

Thomas-282729 (1/1/2011)

---

What is never mentioned in the article is "Scale to what point?". Let me demonstrate: You are using a bigint for a primary key on a table with a...

dpersson-635827 (1/1/2011)

---

You lost me at a few minutes vs. a couple of seconds. If that were truly the case you would have to add an index no matter what,...

TheSQLGuru (1/1/2011)

---

Steve Jones - SSC Editor (12/31/2010)

---

I believe the clustered key is always there. Even if you have a unique NC index, the clustering key must be there so that...

James Stephens (12/31/2010)

---

While I agree in general principle with the gist of the article, I don't think everyone needs to run into work this weekend and take their servers down...

mwenner (12/31/2010)

---

Can someone clarify one sentence in the article:

>>

So you need to keep in mind things like (for SQL Server at least) NonClustered indexes typically contain the data...

I...

Sorry it has taken a while for me to respond today but work has been busier than expected.

First off, thanks to everyone for the compliments. They are truly appreciated. 🙂

Next,...

Gianluca Sartori (12/31/2010)

---

I couldn't find a MySQL instance, I was sure we had one, but I couldn't find it.

I executed the following queries to discover the behaviour of the DB...

Hello Gianluca. Thanks for a great and thorough article :).

This is slightly off topic, but how do...

Bradley Deem (12/29/2010)

---

I think part of this limitation is due to the lack of ability to perform an ORDER BY clause with an aggregate.

Assuming I've implemented it correctly, this could...

David.Poole (12/29/2010)

---

Thanks Solomon, what I was getting at with the Median is that to find the middle value you have to keep track of all the values and my worry...

Adam Machanic (12/29/2010)

---

Have you compared the performance of doing a median calculation w/compression against performing the same calculation using Joe Celko's ROW_NUMBER technique?

http://sqlblog.com/blogs/adam_machanic/archive/2006/12/18/medians-row-numbers-and-performance.aspx

Compression is far from cheap, and I suspect...