Thank you for all your replies. I do not have the perfect solution yet, but definitely a few new ideas. I tried ta add some good indexes on the tables...

Did you really update 9M records in 30 seconds? That is fast. How come? When I...

@Allister: Are you able to make this query without the need for the new columns sec or diff? If I do need to recalculate these column each time I make...

@matt-2: The query was ment as an example. Maybe it was not a good example. The application which will report the information from the database will show the...

:unsure: Sorry if I am mistaking but, this way it looks like I need to recalculate and update 30 Million records each time before I can query the needed time...

Thanks a lot. I am going to test this tomorrow and post the results.

Regards, Tim

No, adding a column is not a problem. What should be concidered is that the amount of io's to this table is pretty huge. Mostly selects and inserts. Not so...

You are right! I am sorry for leaving this information out...

In dbo.Trips I have 2 indexes:

- Unique nonclustered index on ID

- Unique clustered index on Company_ID, Employee_ID, Start_trip

Hi, Thank you Garadin for willing to help me!

A stored procedure is no problem.

Sample data:

--===== Create the test table with

CREATE TABLE dbo.Trips2(

[id] [bigint] NOT NULL,

[Company_ID] [int] NOT NULL,

[Employee_ID] [nvarchar](50)...