Each concept will require a study of new thinking. Your question is a general answer for a specific answer.
You structure and work with CouchDB documents in a different way than with MongoDB documents. In CouchDB, you'll make queries using MapReduce. In MongoDB, you have a flexible query interface similar to an RDBMS.
A repository of key value requires a completely new way of thinking. You must know your query patterns before you can properly structure your content. You do not have an index, so you need to create your own structure.
One blog that provides a lot of NoSql information, http://nosql.mypopescu.com
Update
Riak people also have some interesting questions:
- Will my access pattern be read, heavy, heavy, or balanced?
- What data is collected most? Which of them require more complex conflict resolution?
- How can I find this data type? Which method is most effective?
- How independent / interconnected is this data type with this other data type? Do they belong together?
- How much will I need to do online queries on this data? How fast do I need them to return results?
source share