The explanation is rather silly. Clearly we don't have rounding to the nearest 1/300 sec, because we have numbers like 17 millisec and 103 millisec which clearly are not...

Interestingly enough, SQL 2005 and later allow GO to be commented out even if database compatibility level is set to 80 (ie SQl 2000 compatibility).

Wasn't it fixed in an...

Steve Jones - SSC Editor (2/11/2005)

---

Sorry, it's probably misworded slightly, but basically 2NF from my understanding, involves moving things to lookup tables. So if you have a City repeated multiple...

Interesting. Different people seem to have different numbering (and even different sets).

"shredded" seems a peculiar description for something that is neither more nor less than a clustered index on the nodes of the document :hehe:. Caused me to get it...

I don't like seeing the explanation of QotD being used to plug an expensive textbook when the question is adequately covered in BoL.

Depending on the sort of login he wants to create, he may need ALTER ANY CREDENTIAL permission too.

Well, I learned something new here. Now I know at least one way in which life was better wil SQL 2005 than it is with SQL 2006 R2 -...

Horribly sloppy question where the answer depends on a heap of unstated assumptions.

I've lobbed in my two pennorth. I haven't use linked servers since SQL 2000, so maybe it's not useful. My first though...

Alternatively, give the user who makes the linked server access from server A permission to execute as a (windows) login which is valid login which is valid on server B,...

I think you need to modify the linked server security information on server A.

On the security page of the linked server's properties on server A, tell it what security context...

SQL Kiwi (3/13/2012)

---

L' Eomot Inversé (3/13/2012)

---

That's pretty awful. I hope they fix the bug soon.

If I've understood it correctly, the bug can't happen if isolation level is...

square

SQL Kiwi (3/13/2012)

---

L' Eomot Inversé (3/13/2012)

---

Of course this is absolutely stupid behaviour for the compiler since it is clear that a coalesce invocation where the last argument is not null...