Hi all, I'm wondering how to handle following scenario:
There is a pool of data which can be searched by all users via a web interface. Now some users would like to add own documents, or private fields to existing documents which shall not be accessable by others. Concerning private fields .. I could i.e. prefix user data with an identifier, which would be automatically be prepended as OR clause in the search. Well, it's a hack which I don't like and I don't even know if this will work for phrases ... ... I could widen the schema to add user specific fields, I know there is no performance impact on empty fields, but it still does not feel right to add fields for each customer. Do I have to rebuild the index each time I add a field? Concerning private Documents ... I could add a scope field where a user document holds the user's identifier(ID) or "public", when it's for public view ... I could use distributed search and add a shard per user ... could work, but aint the planned use of distributed search I guess Well, nothing what fits my tummy. This scenario could get even funnier, when customers demand to have a nindividual schema, is this when I need to look at couch or mongo? Maybe Erick Erickson already answered my questions and I should just add fields as I need them while dropping any db knowledge. Could someone please guide me. kindest regards, Manuel
