SOM 😛

With a proper index in place, a TOP(1) subquery can beat the performance of a min(), because the query plan only has to do a one-row seek. I...

Looking at the number of posts you've been involved in lately, and extrapolating from that a certain percentage of BOL lookups, factor in typing time.....

Tell me you have a life...

I still don't understand why the obvious query is unacceptable.

select assetID, min([date])

from t1

where [date] < inceptionDate

group by assetID

Understood and agreed. I'm just glad that we're on the same team.

When all code is posted, and posted correctly, we have a far better chance of getting the answer right on the first try. Please repost everything.

/* Create a...

Paul, please tell me you don't have all that committed to memory. 😉

Paul, thanks for the save. I wasn't even close to thinking about isolation levels last night.

1. Why do you want to avoid the min() function? That's like saying "I want to saw a board in half, but I don't...

Rather than creating an index on the actual table, I instead use an existing index to filter down the records as much as possible, and store the results into a...

Next question: If the objective is to prevent two rows from having the same TableID, why isn't the primary key on [TableID] only? That would put...

The duplicates you mention: Are they simply duplicates of the [TableID], or is the [createdTime] also identical. Since it's a datetime field it would have...

I don't understand what a temporary table would do that a covering index wouldn't do.

If your data isn't read-only and non-volatile, maintaining a set of temporary tables...

I just used your procedure's BEGIN/INSERT/VALUES/COMMIT logic to insert multiple rows with the same GUID.

The trigger worked just as you intended it should.

We are...

Okay... I obviously misunderstood your issue. I agree that a single insert with a VALUES clause like that shouldn't create two rows. (This is where seeing source...