The best would be if it's accessible from python and works as a system service (remote access).
From my experience, MySQL is not capable of doing so, so I'm looking for some NoSQL solutions.
As a NOSQL database you might want to take a look at CouchDB.
However, I would also suggest to look at PostgreSQL, I believe it is a bit more powerful than MySQL. Hope this help.
Regarding NoSQL solutions: CouchDB is an option, but I can also suggest MongoDB, which "ships" with a nice documentation and quite mature APIs (http://api.mongodb.org/) - including Haskell, Erlang, C++, Python, Java, .... . Maybe worth giving a try :)
Regarding MySQL I agree with Paolo that PostgreSQL seems more powerful than MySQL. For details refer to a rather old comparison: http://www.bostongis.com/PrinterFriendly.aspx?content_name=sqlserver2008_postgis_mysql_compare
Thanks Johannes.
I would like to add (it just came up to my mind) that another possible option (since there is the necessity to deal with relationships) might be to resort to a graph database: for example HyperGraphDB (http://www.hypergraphdb.org/index)
I'm playing with some successes with Neo4j (http://www.neo4j.org/) ... I'm quite fed up with SQL, but going to try other suggestions:)
neo4j is a good option (due to property graphs supported), but the suggestion of Paolo - hypergraphdb - goes even one step further, as hypergraphs are supported (nb! an edge may connect any number of vertices). For details just refer to Wikipedia or Google - there are tons of papers around.
Also keep in mind that PostgreSQL can be extended with external functions written in several programming languages like C (and C-like ones), python and so on (take a look here http://www.postgresql.org/docs/9.0/static/xfunc.html). This means that you can use whatever library you can find out there: for example the boost graph library (http://www.boost.org/doc/libs/1_53_0/libs/graph/doc/index.html) which is a prototyped library for graph handling, however, last time that I checked, it does not provide direct support for hypergraphs. Still, if you need hypergraphs you can find a suitable library and write your own extension functions for PostgreSQL.
I support the Neo4j option. There is also some good and easy-to-learn drivers for Python, Scala, Ruby, Javascript, etc. Native support is for Java, but using Cypher, the query language, any of the REST bindings works good for that amount of data.
In the end, neo4j turn out to be too slow either. MySQL works fine, but splitting data over multiple tables. Data is redundant, but access time very fast!
Dear Sir,
with best wishes, why don't you use simple excel file, you could access it using python or any other programming languages.
all the best,
saeid parsa
I think Neo4j and Titan support those amounts of nodes and relationahips and have decent Python clients. OrientDB possibly too, but I don't know about the Python support.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python