TheMOTU (10/26/2010)

---

Guys,

As usual - the amount of difficulty never disappears...

Surrogate keys are excellent for Referential Integrity - small PK / FK.

The default of having PK as a clustering one is...

I think you're right Hugo.

ebay-1087924 (10/26/2010)

---

It seems to be a rather common misunderstanding, actually, especially among DBAs without a background in computer science. Perhaps I should write an article to clear things up....

Agreed.

Our Database - about 200 tables - is about 80% Surrogate/ 20% Natural.

The below is an example where we too prefer the Natural.

In my experience, the only...

wagner crivelini (10/25/2010)

---

Denis

As I said on this forum, I'm also a fan of surrogate keys.

Actually, my goal in this first article is to show the concepts and methods I'm using...

I...

cdesmarais 49673 (10/25/2010)

---

Surrogate keys on pure junction tables are bad, but it seems like alot of people are putting clustered covering keys over both columns, which can be almost as...

I found the article was well done visually and well written with a good background on the pages.

But it seems the article is more about varchar indexes vs integer indexes...

I have found this feature very helpful but could be improved, I would even call it buggy:

- As Microsoft mentions, it does not consider trivial Plans. Why not? A trivial...

Our department eliminated blocking by using 'nolock' table hint. It has worked great, or so we think.

I believe the best option is Snaphot Isolation Level:

http://msdn.microsoft.com/en-us/library/ms345124(SQL.90).aspx

Then readers don't block writers and...

This will ensure the Deadlocks get written to the SQL Server Log file:

DBCC TRACEON(1222,-1)

And you should consider adding these to the Start-up Parameters.

(2)He also says that "You can DELETE any row that will not violate a constraint, while leaving the foreign key or any other contraint in place.".

Does he mean that,...

This is an interesting one.

What Gila Monster mentioned makes most sense to me.

According to this, Truncate table acquires the Sch-M (schema modification) lock:

http://weblogs.sqlteam.com/mladenp/archive/2007/10/03/SQL-Server-Why-is-TRUNCATE-TABLE-a-DDL-and-not.aspx

Whereas a delete won't acquire this lock. So...