Don't just copy and paste, read the article and understand...

I also knew the fact the question was looking to test, but fell foul of the start at zero like so many others.

I personally think the question would have been...

Evil Kraig F (8/22/2012)

---

Just on a side note, be careful with indexed views, particularly top 10 views and the like. On paper it doesn't look like they'll cost much...

Certainly best practice for ad-hoc batches, but I would say it is merely...

The NOT EXISTS...INTERSECT thing works in SQL Server 2005 of...

Another example, using APPLY syntax:

DECLARE @T AS TABLE

(

pk integer PRIMARY KEY,

col1 integer NULL,

...

sqlfriends (8/22/2012)

---

Because I have quite a few <> comparisons in my sql, ( in my post it is just a part of them), if I write something like :

...

You make some good points, Joe, but I fear the "proprietary module" allows for limited rewrite & redesign opportunities.

Yes. The point is that each child record must...

Col1 needs to have a PRIMARY KEY, UNIQUE constraint, or (less acceptably) a UNIQUE index defined on it. Something needs to enforce uniqueness on col1 alone. The composite...

ChrisM@Work (8/22/2012)

---

Updating a clustered index key causes a delete/insert to the table and updates any non-clustered indexes to reflect the new key. And whatever else. That's a lot of (unnecessary)...

Using WITH RECOMPILE on the procedure...

Sergiy,

I've been trying to work out why you're getting so excited about all this. Perhaps you are wondering whether an update to the same value is actually written to...

ChrisM@Work (8/16/2012)

---

It's interesting that in only the cheapest case has update to self been optimized out. Perhaps checking for change in a LOB column is more expensive than optimizing out...

Sergiy (8/16/2012)

---

Now it's not a heap anymore.

You may try it for yourself - still no dirty pages.

Sorry Paul.

It's different only when PK is declared on (Name).

Then it goes like Paul...