>>Exactly - I was going to tell that it's not going to work this way because of multiple inserts possibility.
>>
>>But would this work?
>>

>>    INSERT INTO ProductFileData
>>    (ProductFileColumnId,ProductFileRowId,ProductFileId)
>>    SELECT INSERTED.Id,NewID(),INSERTED.ProductFileId
>>    FROM INSERTED INNER JOIN ProductFileData ON ProductFileData.ProductFileId = INSERTED.ProductFileId

>>
>>Or why not let ProductFileRowID field to have default NewID() and then you don't need to include it in insert?
>
>Actually what I posted above is complete rubbish. The ProductFileRowId has to be based on existing unique values for existing rows.
>Using NEWID() is completely wrong. Using a default would be worse because all the rows would have different Ids. Let me try again....

Can you post the problem definition again with the table creation scripts and some data?