On 27/06/14 08:50, Paulo Picota wrote:
Hello friends =)
Im using ARQ to do some sparql queries to a OWL ontolgy for my Master
Thesis. One thing I need for my thesis is the order O(n) of the graph
pattern matching algorithm that is used to resolve the querie after its
final transformations and optimization steps.
(http://jena.apache.org/documentation/query/arq-query-eval.html)
I have been looking for a while now and I cant find any reference to what
algorithm is been used.
You mean the basic graph pattern matcher?
In the released code, it is a form of index-join done by substituting
results from earlier parts of the pattern repeated into the later parts.
It's not O(n) for whatever the n is measuring.
It is constant memory - one of the advantages of the implementation is
that it does not consume increased working space as the data increases.
For patterns that have moderate
There is no reason it has to be this algorithms - there is code
elsewhere that uses hash joins, merge joins or sorted merge joins.
Merge joins are useful in TDB (if you run with a different index set).
if there is a paper or any one has the information about the order of the
algorithm, its medium and worst case scenario, or maybe some documentation
I could read about this, it will be a great help.
The code is good place to look.
Andy
thanks So much,
Best regards =)
