Saritha... it would be really nice if you shared how you solved it. 😉

No... if you're talking about simple duration measurements, it's not useless... you've just gotta make sure they're either both "hot" or both "cold" 😛

But, in the face of blocking, looking...

Heh... Thanks, Noel... I keep forgetting to check the date on the original post...

Oh well... keeps my hand at typing and I have seen it where the OP actually...

The update method works well, in all cases... and it does NOT depend on a Clustered Index... just SOME index.

You just need to settle down, maybe have a beer, and...

Hey folks... I had reason to revisit this post to demo to someone else... I don't know what change was recently made to the 2k server I work on but...

Linda Robinson (10/9/2007)

---

Jeff, I agree there should be a better way, but since the people who own the application like their datapump (I've talked to their tech support), I'm stuck...

You certainly did take that one out of context...

The real problem is that you're trying to solve two different problems at once... divide and conquer... the code will not only be faster performance wise, it will also be...

I'm by no means making fun of you, Linda... but the polar ice caps are retreating faster than "200 rows per minute" and it's why I don't like things like...

Matt is correct about needing the CASE... but I'm not sure why we need the #Summary table, at all...

First, a brief lesson in creating massive amounts of test data in...

Perfect... I'll be back...

Sorry... that just doesn't help me... all of your examples show only one date... would you please do as I ask so I can help? Please change the following...

To add to that... are there Primary Keys on these tables? Is there a clustered index? Are the tables in a constant state of flux do to inserts,...

If it's the "Estimated" Execution Plan... pay those percentages no mind... it "lies" because it will frequently miss things. For example, only the first pass of a triangular join...

Heh... thanks for the "apple" 😛