Prior to 2005? Suffered.

There really wasn't an easy way to do it back then. The best approach was to have a natural key on the table then you wouldn't have...

You mean like "You're right and I'm wrong" as I just recently posted?

😛

Sorry, you're right. It's leakage from the old days. SCOPE_IDENTITY specifically deals well with triggers. My bad.

I still like the OUTPUT approach becuase it works the same with one row...

It doesn't work on 2005. Period. Full Stop.

It doesn't work terribly well on 2008.

Get Red Gate SQL Prompt. It works great on both and in Visual Studio (although it's a...

I assume you already updated the statistics with a full scan?

After doing that... Can you post the execution plan & query? Some functions that worked well enough in 2k don't...

As good a forum as any. This is a huge topic, but I'll try to...

If you're scripting database creation, two areas are going to cause you problems in terms of time. First, however big the database you create determines, on a given system, how...

This sort of implies that you're only using the GUI to create tables. You can change the schema for the object created in the gui by accessing the properties window,...

It really depends on your business needs. If you need to keep a history of the individual call itself, your method won't work. If all you care about is the...

Matt's nailed it. If you went for the identity value on the table, you could get someone else's insert value. OUTPUT pretty much guarantees what you're getting. Still, in most...

Huh... I'm kind of amazed that works. I just tested it. Not only can you do it, but it doesn't affect inserts or updates. Strange.

No, I can't imagine why you...

You can also open up the Properties window and set it there.

Best of all, try using TSQL. I think it's a heck of a lot easier to create tables there...

For a single value from OUTPUT, I'm pretty sure it's free.

Where you might see a performance hit (or benefit, depending on how you define it) is when you define...

Garadin (11/25/2008)

---

The note about clustered indexing brings to mind something I came across in a section mockup yesterday. When making Foreign key reference tables (IE. Tying an employee to...

That one was a hoot. The original problem definition was what, four lines. It grew to almost four pages of code by the time the guy was done adding things.