Document-oriented databases vs plain text files

Question

I'm working in a Python program which has to access data that is currently stored in plain text files. Each file represents a cluster of data points that will be accessed together. I don't need to support different queries, the only thing I need is to retrieve and copy to memory cluster of data as fast as possible.

I'm wondering if maybe a document oriented database could work better than my current text file approach. In particular, I would like to know if the seek time and transfer speed are the same in document-oriented DBs that in files.

Should I switch to a document-oriented database or stay with the plain file?

Answer 1

You probably should check MongoDB according to this answers and MongoDB documentation it should have much better performance and more advantages over plain texts like easier backups and it has also nice API in Python.

Answer 2

A DODB sounds like a much more reliable and professional solution. Besides you can add stored procedures thinking in the future and besides most databases offer text search capabilities. Backups are also easier, instead of using an incremental tar command, you can use the native DB backup tools.

I'm fan of CouchDB and you can add RESTful calls in a "transparent" way to it with JSON as the default response.