Thank you all for your reply, i think you're right, and the problem is probably during to blocking issues.

I'm working on this, and i'll update about the solution if the...

Thank you very much, that's really what i need.

I searched for hours i did not find anything about it.

Best Regards.

If you can change the schema as Stephanie Giovannini said, do that, because it's really a wonderfull schema design.

Use views instead of CTE, and you'll get the same result.

sorrowfully, currently i don't have SQL Server 2000 installed on my PC, so i can't test it.

Test it and tell...

yes you can in many ways, but you have to provide us with a sample record first.

Sure, if you delete from the lowest level up to the highest one, you do not need any more to disable any constraint or use the cascade option.

First of all, you have a mistake in your design in the first table, because as the relation between the phone and the customer tables are many to many, then...

I think you're making an auditing system for your database.

if you put the trigger on the first table, you have to worry about two points:

- Generally speaking, and according...

you can make usage of DELETED table, which you can access inside the DELETE trigger.

it's just a JOIN statement

I think you want the trigger to be on the third table,...

i didn't test such algorithm for a 40 million record, but i think that using cursors is the worst solution for such a problem.

you have to talk a lot about...

really the equations are not pre-determined, and they're written by the customer for each record.

So i can't use CASE expression.

Thank you anyway for your participation.

yes, and i think for the delete operation your solution is better.

I think you can use ranking functions and assign a row number for each record, and then you can get the last row.

SELECTTOP(1) QuestionId,

ROW_NUMBER() OVER (ORDER BY QuestionId DESC)...

Thank you a lot, really i'm using this technique with NULL values, but i made it more complicated for me with this condition.

Thank you very much, you saved me a...

Great, it's working

thanks a lot.