>>>Last question... Which is the common way adopted by everybody...
>>>
>>
>>Vinod,
>>
>>There is not one single "best" way to do that. I usually have a surrogate primary key for each table, that is a key that bares no relation to the data
>>and ensure uniqness
>>
>>Arnon
>
>can u pls give me a example...

if for example you have a table of people
you can set a person's ID number as a primary key and that would be a natural key - you can however add an additional field a numerator that will be incremented without relation to the data itself and set that as primary key - this would be a surrogate key

Arnon