NoSQL Database Structure?

Question

I am making a makeshift non relational database management system. What I am actually implementing may defeat the purpose of why you would use a non relational database, but this is for a school project.

QUESTION: Below is a list of more details. I want to know what else would I need to create or eliminate from my inital process to implement an example of a NoSQL document database.

That database will work hand in hand with a web application that will store a name, zipcode, message, subject tag (fixed amount of options), and a timestamp. I want to be able to index data, update a feed based on the time stamp or based on the subject tag or the zipcode range.

I plan to make a document store with json. Want to implement an example of a map reduce (more info on this would be great), and a psuedo replication factor. The language of choice is Java.

Thank you ahead of time, and I will be responding to all comments and posts.

EDIT: I am not interested in using another database at the moment. I am only interested in creating one. My question just asks what other features should I implement that I have not already stated. That is all.

Answer 1

You can consider using MongoDB , and this could be a good exercise to learn a NoSQL database. It uses BSON as the storage format, and has a very simple query language . You can write complex queries (including mapreduce) in Javascript. Since you want to program in Java, it has a Java Driver , and if you want, you can use an ORM like Morphia to make it even simpler. It supports indices and in-place updates/upserts for your update requirements. You can replicate it by setting up a Master-Slave replication in no time, just by flipping the startup argument. I would like to know more about your requirements before answering the map-reduce question though.

Feel free to read about Mongo and Morphia, I am sure you'd find it both interesting and exciting to dive in.