Getting Started With SPARQL, Sandboxing

I'm a firm believer in learning a platform or technology by playing with it. So I've kind of glanced at SPARQL and thus far haven't really gotten to into it because it just ... looks daunting (as does everything at first).

I'd like an environment with a couple of documents that I can cross reference and query. My temptation is to do SQL like things such as joining a city to it's home country - but maybe that's the wrong frame of mind. I've looked at the examples on dbpedia and right of the bat there's question marks, 5 namespaces, angle brackets and other hoopla.

http://dbpedia.org/snorql开发者_如何学Go/ looks like a good place to play.

Anyone have some experiences to share? Any resources in particular that really launched your learning?

I've heard that the ARQ SPARQL tutorial is a useful start. As you've found, dbpedia is a daunting data set to start with, but sparql.org also provides a very simple dataset to try querying online.

I agree with you. So, I have put together a starter kit to help with starting an initial data set. It's still in the very early "alpha" stage, but I have been able to use it to create a set of triples to work from. https://github.com/barobba/RDF-SPARQL-Starter-Kit

Also, here is a related post I just made on StackOverflow: How to get started with RDF? (Particularly for relational database developers?)

As comment_bot says the ARQ SPARQL tutorial is a good place to start.

If you really want to understand it then you will need to read the full SPARQL specification which has lots of examples and explains it in a lot of detail. If you just want to use it then you shouldn't worry about section 12 onwards as that goes into in-depth technical details of the formal definition of SPARQL and how you might go about implementing it.

But also like you say you do need to just play around with queries - again comment_bot mentions sparql.org which is a handy sandbox

Also be aware that currently there is a Working Group actively working on SPARQL 1.1 which will add in a number of useful features such as aggregates and negation which are either missing or hard to do under the existing specification.

You may also want to keep an eye on SemanticOverflow which is the Semantic Web equivalent of SO and quite a few of the people involved in writing the SPARQL specification visit that site so if you have an obscure/complex SPARQL question it's the place to ask