Thanks. That answers my question.

So, if I understand correctly, the execution plan costs should be used as a good indication of which parts of a query are ripe for optimisation...

Thanks for the response.

How can the costs be estimates in the actual execution plan?

I thought the actual plan was the plan it used for the query?

I did some further testing last night and the delay in the alter table alter column statement only occurs when a non-clustered primary key exists on the table.

It completes in...

....but when it's just the clustered index with no non-clustered indexes it's mili-seconds rather than 10s of minutes? This would suggest it's just updating the meta data.

Hi,

Thanks for your response.

I now understand the reason for the time taken to drop a clustered index.

I am still confused regarding the answer to Q2. If there are no non-clustered...