Thanks Ravindra,
I think you're right when you say it will not perform any data consistency checks. But then why the name "Key Attribute".
Also we can have joins in analytic view which are not based on key attributes. So how does the optimization take place.
Do you have some document which explains this concept in detail? I am still not clear after your explanation.
------------
I did some R&D myself to get some clarity on this concept. I found that the Key Attribute is required to identify the Central Table, So the key attributes can come from only one table in an attribute view with multiple tables.
What I did was to turn off the client side validations and then try to activate an attribute view
1. Without any Key Attr., It gave an server side error about 'No Central Table found'.
2. Next with Key Attribute from one table, it activated in both the cases. CUID and CID as expected by now.
3. Last with CID from one table and CUID from other. It gave the same Server side error about Central Table.
So if I make a attribute as key what makes it special apart from identifying the Central Table. If that was the purpose then why don't we have an option in Attribute View for just directly selecting the central table?
Also the joins in analytic views can be made based on an attribute coming from non central table?
I am very confused. Please help
Thanks & Regards,
Shreepad Patil