The queries don't do the same thing. Lucene queries aren't exact match.
If you are comparing exact match against exact match, for a database
that has warmed up, the triple pattern lookup is probably faster.
If the query is only doing it once, then it would be hard to time the
difference.
If the database is cold (caches haven't stabilised, JIT hasn't run, etc
etc), it's anyone's guess.
Andy
On 24/07/2019 10:45, 刑天 wrote:
For example, I have using textindexer to index a lot of SPO. and stored
rdf:label predicate's value in text index.
And when I use:
PREFIX geo: <http://www.opengis.net/ont/geosparql#>
PREFIX text: <http://jena.apache.org/text#>
SELECT ?subj
WHERE{
?subj text:query (rdfs:label "Some Text")
}
Is it faster than:
PREFIX geo: <http://www.opengis.net/ont/geosparql#>
PREFIX text: <http://jena.apache.org/text#>
SELECT ?subj
WHERE{
?subj rdfs:label "Some Text"
}
