Just to play devil's advocate for a minute, I think there are some indicators of past performance outside of work that could indicate future performance on the job. For...

peter (2/19/2008)

---

Good thread - nice work, everyone.

Also, I can't recall how the indexing issues related to the SQL 2000 generated uuids have translated into 2005...I vaguely recall some ideas about...

I don't know about whether outside performance is an indicator of performance at work, but I definitely have the opposite problem that you have with productivity. I occasionally work...

Brandie Tarvin (2/14/2008)

---

My apologies for the poorly worded question, but I'm still glad I posted it. I'm learning all sorts of things from this discussion. The primary thing...

Matt Miller (2/18/2008)

---

Brandie (who put the question in to Steve) mentioned she got it out of one of the MS Press training books...God only know where they might have picked...

Matt Miller (2/18/2008)

---

From what it looks like (based on the logical reads you see in Profiler), if you go to the trouble of marking it as deterministic (by using the...

What if your UDF is a scalar detereministic UDF with a constant value passed in? How many times is this UDF executed:

CREATE FUNCTION dbo.AbsoluteValue (@i INT)

RETURNS INT

WITH SCHEMABINDING

AS

BEGIN

...

On the 1,000,000 row example with clustered and nonclustered index I get 16.7 seconds with this one:

DBCC FREEPROCCACHE

DBCC DROPCLEANBUFFERS

GO

SET STATISTICS TIME ON

GO

SELECT d.SomeID, dbo.fnConcatTest(d.SomeID) AS CSVString

FROM (

...

That you did! Pascal is the most delightful pseudo-academic with a grudge that I've ever had the pleasure of...

Matt Miller (2/13/2008)

---

Stop - you're starting to sound like our "friend" Cimode's position from the infamous epic flame/religious war experience, NULL versus Null?[/url]

for what it's worth - that thread...

chris.compton (2/12/2008)

---

Okay, with the risk of sounding more ignorant than I am, I'll admit that I get a little lost.

I've been programming for almost two decades, and the terminology I...

Cary Hower (2/12/2008)

---

This is really a separate topic from identities, but I feel that PKs should be clustered because they are so frequently used in joins. It seems that join...

SteveRowland (2/11/2008)

---

Without an SID, how would you store multiple addresses for a person, and be able to modify a single address, without using every field in the address? Would you...

Matt Miller (2/11/2008)

---

That's an interesting thought, but there are times when you might get the same data (say - you're tracking outcomes of consecutive coin tosses). There really might...

jaholbrook (2/11/2008)

---

identity columns are like the fast food of the database world.

..."Super Size Me" ring a bell.

Basically a lack of understanding on the upfront work involved in modeling. Does...