--- Update ---
Thanks for the comments and the help so far. And I apologize for not specifying the question further. I have updated the question below.
--- Update ---
Currently I am asked to develop a database structure for a pretty big amount of data. I am debating between implementing a graph database instead of a "normal" relation database, and was wondering what the disadvantages are if the data does not necessarily contain any relationships? Could separate, unconnected, nodes be utilized the same way as a table / row in a table?
I am asking this because as of right now there are is no need for relationships but I am trying to future proof the database (with the anticipation of relationships) to expand the abilities of the data. If it is of any help, I am debating between OrientDB/Neo4j or mySQL/postgreSQL.
Let say we have a database full of stocks. Stock can be bought and sold at pretty much any time / day by anyone (as long as the market is open). Now this database could be a normal relation database:
Table 1: IDs | Products | Prices | Sizes | Dates. But could potentially also be organized as a relationship database
Node 1: Stock A | Node 2: Stock B.
If I just use the database for storing stock information, it seems to me a normal database will be better. But is this true? Will it negatively affect me working with a relationship database / will I be better off working with a normal database? Are there disadvantage to organize my data in nodes, instead of rows?
A picture to illustrate it all: http://tinypic.com/r/2zss8s3/6.
The database will mainly be used for products bought and sold similar like stocks, however they also have other information such as locations attached to them. I am trying to anticipate the implementation of features such as recommending products or even the prediction of next products to be bought by somebody.
Most of the queries from the database will be per product.
Pulled from the database: 100 to 1000 times a day.
Pushed to the database: 20.000 a day.
An added question which might reveal a disadvantage:
How easy will it be to go from relation-to-graph or from graph-to-relation. Any lock-in dangerous?
Thanks for all the help, comments have been great so far! R