Hi all, I am modelling a table for a shopping site where we store products for customers and their data in json. Max prods for a customer is 10k.
>>We initially designed this table with the architecture below: cust_prods(cust_id bigint PK, prod_id bigint CK, prod_data text). cust_id is partition key, prod_id is clustering key and combination of these is unique. >>Now, we have a requirement to store one ordering column "prod_order (bigint)" which is pre-calculated from service end after complex calculation. Further we now want to want to display products in pagination (100 per page in order of prod_order) Please suggest the new architecture for the table "cust_prods" so that : 1. We could be still able to update a single prod info based on cust_id + prod_id 2. We could store data in order of "prod_order" so that fetching in limit(pagination) becomes easy. Note : prod_order could change for a product frequently Regards, Ashish