Paul White NZ (4/1/2010)

---

Hugo Kornelis (3/31/2010)

---

...I submitted a bug report... a case I constructed, based on this discussion, where COALESCE(..., 1) may erroneously return NULL (which should of course never...

Paul White NZ (4/1/2010)

---

On the subject of IAM-driven scans. You are, of course, absolutely right about the conditions for an IAM scan on a table with a clustered index...

You are right! Thanks for pointing it out. The sixth option should have read "A clustered index exists ..." instead of...

Yes, I noticed that...

The query uses [font="Comic Sans MS"]IsNull(@MaleNames + ', ','') + [name][/font].

When @MaleNames is NULL, [font="Comic Sans MS"]@MaleNames + ', '[/font] is NULL as well, so the ISNULL kicks in and...

Hugo Kornelis (3/31/2010)

---

my gut feeling is that, if the same data is used multiple times in a query, it should use the same value, regardless of concurrent modifications.

(And, of course,...

Paul White NZ (3/31/2010)

---

Did you read the comments in Alex's blog entry (including mine) or the previous article?

My point there, as...

sknox (3/31/2010)

---

If each statement were indeed treated as its own transaction, then Hugo's query should always return the same results in all three values, because it's a single statement (even...

Paul White NZ (3/31/2010)

---

I know you know this! It has...

Paul White NZ (3/31/2010)

---

Hugo Kornelis (3/31/2010)

---

I would agree that, depending on transaction isolation level and @@TRANCOUNT, the results of the three statements might be different on a busy server. But...

Peter Trast (3/31/2010)

---

Tom.Thomson (3/31/2010)

---

Peter Trast (3/30/2010)

---

You don't like T-SQL much, do you? 🙂 Seems to have a bad effect on the...

Paul White NZ (3/31/2010)

---

Hugo Kornelis (3/31/2010)

---

I disagree. If the query includes some non-deterministic code (such as NEWID()), there might be room for argument. But if the query in itself is...

Paul White NZ (3/31/2010)

---

Hugo Kornelis (3/31/2010)

---

So if any of the arguments is a subquery, the implementation if COALESCE((subquery), ...) becomes CASE WHEN (subquery) IS NOT NULL THEN (subquery) ... ELSE...

James Lean (3/31/2010)

---

Do you have a supporting link or further information on the performance difference...

Paul White NZ (3/30/2010)

---

Me either! I tried indexes,...