Thanks a lot for the help. Can you please tell me the name of the file where I can find these functions?
On Mon, Nov 25, 2013 at 8:40 PM, Andy Seaborne <[email protected]> wrote: > On 25/11/13 14:23, Rob Vesse wrote: > >> Maria >> >> Yes you can do this, in ARQ the iterator which backs the ResultSet or >> Iterator<Triple> returned from the appropriate QueryExecution.execX() >> calls represents the plan and iterating over it causes the actual >> execution to occur and is thus the execution time I.e. >> >> // Assume we have a query and a dataset ready to go >> QueryExecution qe = QueryExecutionFactory.create(query, dataset); >> >> // Starting to execute the query only actually triggers plan calculation >> // Time around this to get the plan compilation time >> ResultSet results = qe.execSelect(); >> >> // Iterating the results actually executes the query >> // Time around the while loop to get execution time >> while (results.hasNext() { >> results.next(); >> } >> > > That's the best there is. It's not as clear cut as completely separate > plan and execute stages. Only the high-level optimizations are measurable - > TDB specific low level optimizations happen during execution. > > The major cost for TDB is whether the file caches are hot or cold. > > Andy > > > >> Rob >> >> p.s. Please DO NOT email/CC developers with your question as well as >> posting to the list or re-post your question multiple times, developers >> are all volunteers and will answer your questions as and when they have >> the spare time. Also as Andy points out people are in potentially vastly >> difference timezones to you so you should always allow at least 24 hours >> for a response if not longer. >> >> Emailing people directly as well as the list or multiple posting only >> serves to annoy people and actually reduces the chance that people will >> respond to your questions. It also subverts the purpose of the list which >> is to allow any knowledgeable user the chance to respond and share their >> knowledge with the wider community. >> >> On 25/11/2013 08:08, "Maria Jackson" <[email protected]> wrote: >> >> I need to know these query times as I am benchmarking Jena TDB against >>> postgreSQL. With postgresql I am able to separately retrieve the query >>> execution time after the plan has been generated by postgresql using >>> \timing. It would be great if Jena could also give me the time it takes >>> to >>> retrieve the results after the plan has been generated separately by Jena >>> as that would help me take a one to one comparison. >>> >>> I'll be really greatful to you for the guidance. >>> >>> >>> On Mon, Nov 25, 2013 at 5:48 AM, Maria Jackson >>> <[email protected]>wrote: >>> >>> One can find time of query execution in Jena from tdbquery using --time. >>>> But does this --time include the plan generation time. If yes, is it >>>> possible to find: >>>> (plan generation time) and (time to retrieve the results after the >>>> plan >>>> has been generated separately in Jena)? >>>> >>>> >> >> >> >> >
