Nevyn (7/8/2011)

---

...but it looks like PIVOT and UNPIVOT don't slow actual data lookup, but are massive CPU hogs in formatting the returned data. So when he threw 16 cores...

16-core test machine:

---------------------------------------------------------------------------

build temp tables

---------------------------------------------------------------------------

String manipulation (dups,...

opc.three (7/8/2011)

---

It looks like your solution is removing dups. I modified it to leave them in which should make it a bit more efficient..hopefully I did not mess it up....

Yay! Database Developers FTW!

It strikes me that the elapsed time of the tests is limited by the time it takes to display the results - consider using SELECT INTO to write the results...

http://www.ssrstips.com (7/6/2011)

---

I thought that the CTE (it's a non-recursive one) constructed an in-memory table, and that as both the select and count worked on the same in-memory table, there'd work...

Badly?

Understood, and agreed 🙂

GilaMonster (7/7/2011)

---

I've worked with a lot of them. When all queries against a table filter on the date (inserted date, transaction date or similar), the date column is a good...

Sure.

http://www.sqlservercentral.com/Forums/FindPost1013407.aspx

Jack Corbett (7/7/2011)

---

...I'm still debating attending the Summit. I think it is a worthwhile investment because of the content and the networking, but I have to decide is I...

mister.magoo (7/7/2011)

---

With the PRIMARY KEY defined, the query plan shows a SORT DISTINCT which results in 1 row from the join of the two tables, and seemingly this result contains...

Hi John,

There was an interesting discussion on Connect over Hugo Kornelis' request to deprecate UPDATE...FROM:

http://connect.microsoft.com/SQLServer/feedback/details/332437/deprecate-update-from-and-delete-from

There's a bit more going on than that, but essentially yes.

Interestingly,...

Please see:

http://www.simple-talk.com/sql/learn-sql-server/sql-server-index-basics/