As written every time it runs it does one or more selects from inserted and 14 inserts

If  update(blz)

select * from inserted

INSERT INTO [eGecko3_Tr].[dbo].[Änderungen_Spender] (Spender_ID, [Name], Vorname, Telefon, [Straße], Land, PLZ,...

If I'm going to do an insert across linked servers, it looks like this

Insert into LocalDB.dbo.tbl (<fieldlist&gt

SELECT <fieldlist> FROM RemoteServer.RemoteDB.dbo.OtherTbl

Be very careful of select......

Greg. It's quite irrelevent where in the column order the primary key is. I assume that your primary key is clustered (since it's the default)

The data will be stored on...

Regarding the 27 million rows... No cross joins, no bad connections (local machine to local machine), Table we were pushing into had a clustered index. TempDB is free to grow,...

Select date, category, count(*) AS cnt from MyTable

Group By date,category

UNION ALL

Select date, 'Target' AS category, 30 AS Cnt from MyTable

Group By date

That's one solution. I'm sure there are others.

Try copying those 800000 records to another server without DTS. Select ... Into ... across a linked server is painful to say the least.

Also the 'school-of-hard-knocks code' can give problems...

This one will also work.

SELECT * FROM TableA LEFT JOIN TableB ON TableA.ID=TableB.ID

WHERE TableB.ID IS NULL.

Well you start by firing the DBA who let a developer make changes directly in production.

Code moves from dev to QA to Prod, not the other way around.

If you use DTS (which can be very, very fast in the right conditions) specify a query instead of a table and add the with (nolock) clause. That will ensure...

How did you copy the database from production?

Second the question of what exacly you did with the indexes.

How big is the table in question?

It is indeed acceptable for the execution plans to differ. On my production server...

I do a lot of query optimisation at the moment.

I love the feeling that accompanies getting a 5 minute query run in 5 seconds

The odd occation when a user...

Understanding locking is essential if you find yourself administering a heavily-used database (or one with very bad queries and indexes)

Understanding the basics will take you a long way.

the main...

The best choice for a clustered index is a column that is narrow, evver increasing and unique. The closer you get to that, the better. The PK is not always...

The questions you'll get in the exam are selected at random from a large pool. Hence you may or may not get any questions on locks.

What's the problem you have...