1) http://sqlblog.com/blogs/hugo_kornelis/archive/2007/11/28/poor-men-see-sharp-ndash-more-cursor-optimization.aspx. Don't forget to examine the original post mentioned at the top.

2) to Joe: there are several implementations for regex that developers (both XP-based and CLR-based) can...

Yep... already admitted that. CLR's are better at most RegEx and file handling. With those two exceptions (so far), and with only rare exception, people can usually at...

Sorry, posted...

Jeff Moden (1/17/2009)

---

Heh, ok, you said it,...

Jeff Moden (1/16/2009)

---

TheSQLGuru (1/16/2009)

---

Tempdb constraints come to mind here. 🙂

I sure do wish someone with CLR knowhow (and some free time) would take the opportunity to do some CLR...

Tempdb constraints come to mind here. 🙂

I sure do wish someone with CLR knowhow (and some free time) would take the opportunity to do some CLR code you could...

Jeff, it strikes me that pretty much all of your examples here and in other threads have a clustered index on the key of concern. What happens when it...

Jeff:

1) For your follow-on articles, when you demonstrate the 'tsql trick' solution, please make sure you caveat the hell out of it about how and why you can get incorrect...

One other thing to check: SSMS has a specific set of SET statements that it executes on connection and whatever API Peoplesoft uses to connect has another. Check...

Marios Philippopoulos (1/8/2009)

---

TheSQLGuru (1/8/2009)

---

It is my belief that even if you rewrite the trigger to only do the audit update on rows where something is different between inserted and deleted...

Mark, do this - FORCE sql server to use the two indexes you think it should use (from mytable WITH (INDEX=myNCindex)...). Now show the estimated query plan for that...

It is my belief that even if you rewrite the trigger to only do the audit update on rows where something is different between inserted and deleted the trigger will...

I think there are PLENTY of common (and uncommon) ways for SQL Server based apps to perform suboptimally. 🙂 Plus, just because some schmo like me says something...

Peter, I couldn't really follow the code you had. Was it supposed to be just separate batch files or was some/all of it supposed to be wrapped up in...

Which is why I was saying it should be decompiled down to it's logical equivalent, which will do the right thing:

create table #t (fname varchar(20))

insert #t (fname)

select top 100 'asdf'

from...