Flo, are you getting your reads from SET STATISTICS IO ON? IIRC UDFs don't report IO correctly in this manner. Try getting them with Profiler.

Ashish Pathak (5/1/2009)

---

I have gone through the entore thread. Could you please tell me how my solution is so suboptimal. Will it not work in any condition or will it...

Paul White (4/30/2009)

---

Unless you use the solution posted...

Goldie Graber (4/30/2009)

---

If your dates have...

Ashish Pathak (4/30/2009)

---

Posted comment on Comparision of Dates in SQL looks fine. But as written above, it may not be the best solution in case of using indexes. Well... while...

Alejandro Pelc (4/29/2009)

---

I agree with TheSQLGuru about the lack of information to correctly guide ritesh, but also think that tempdb comments are important because of the same reason: didn't provide...

You didn't give us nearly enough information to guide you here. If you want the highest inserts per second, simply have a heap table with NO indexes whatsoever. ...

when using comparison tools one is most often comparing the entire table anyway, so index scan/seek issues...

rob.lobbe (4/28/2009)

---

TheSQLGuru (4/28/2009)

---

Well if you change the code - ie the data type - of course...

craig.lovegren (4/28/2009)

---

rob.lobbe (4/28/2009)

---

Not sure why dates cause so much trouble...

declare

@date datetime,

@today datetime,

@tonight datetime

select

@date = getdate(),

@today = convert(datetime,convert(int,@date)),--use smallint for smalldatetime

...

Josie (4/28/2009)

---

the easiest way to extract date ranges without having to worry about the time parts is to subtract 1 day from the bottom of the range and add 1...

Kurt W. Zimmerman (4/28/2009)

---

My approach is quite simple. Often times I am passing in a date range into a sproc for selection criteria. What I do is the...

Dallas Martin (4/28/2009)

---

Hmm... You may suffer a huge performance hit if you use DATEDIFF().

SQL SVR may have to perform a full table scan. Not exactly what one

desires in a...

A very minor point for Number/Tally tables is to create the clustered index after population and to specify 100% fill factor. This will ensure minimal IO for subsequent use....

What an absolutely amazing thread this is! Well done to all involved. I have learned a good bit, and have some new tools in my bag of tricks!...