“For the ‘UserPhoto’ table, the ‘key’ would be ‘UserID’ and the ‘value’ would be the list of ‘PhotoIDs’ the user owns, stored in different columns. We will have a similar scheme for the ‘UserFollow’ table.”
Does that mean pk is userID and value is a list of userID the user follows?
And in the pic, UserID1 and UserID2 are both primary keys. I think we there is only one primary key allowed in db.
I would also like to know the answer to this
Cassandra for relation storage confuses me as well. If I understand it right, in one row we store users that are related to some exact user id (for example, followers that follow some user). In this case, to get a list of followers we need to retrieve the whole row and this doesn’t look like a correct usage of columnar DB since operation on the whole row is relatively slow for columnar DB. Am I missing anything?
UPD. Nevermind, I misunderstood an idea. Each column is a set of followers that are related to some user.
I guess the schema is for SQL database, and they choose NoSQL later instead. I am still so confused about which 1 is better over the other. Without using joins in SQL db, we still can use it. SQL & NoSQL both can be horizontally scaled nowadays. I just don’t know how to respond to interviewer when asked the pros and cons.