Usman Butt (9/17/2012)

---

SQL Kiwi (9/17/2012)

---

In any case, the assumption must be that TRY_* will only be used where necessary, and (relatively) slow as the intrinsic functions may be, it will...

Lisa Slater Nicholls (9/17/2012)

---

However... if they implemented the TRY_* functions via an internal error catch as I said before I think there might be a performance penalty for using it....

I think I have read all the comments (from all the times this article has been published) but I haven't seen anyone mention the new T-SQL conversion functions in SQL...

Depending on jet-lag status, I might pop in.

There's a 2005 forum? 😉

I would go with 'proud'...

You can check which pages of sys.sysobjvalues are in the buffer pool and view the page contents using DBCC...

Divine Flame (9/17/2012)

---

Thanks for the feedback Paul. Many of us (including me until I didn’t...

Good question, I guess, though I would have preferred it without the INIT trap.

L' Eomot Inversé (9/16/2012)

---

If the query processor figures out that it is not going to update the row (this is decided on the basis of where clause and on clause)...

I chose a literal string rather than...

Sergiy (9/16/2012)

---

I figured out why you built the test this way and I have no argument about the point:

comparing binary content of pages is and will always be more effective...

Sergiy (9/16/2012)

---

I don't think so.

Even if no rows are updated the tables are still created.

SQL Server still allocates pages for the tables, names all the columns, etc.

They have nothing in...

Sergiy (9/16/2012)

---

Actually, you did. Trigger is executed after UPDATE. If SQL Server figures out that the values are the same and does not proceed with writing to the pages then...

Sergiy, I should add that the fun in the example was in working out *why* that particular case gave the performance results it did. With a long string of...