Hi Fabian, this is now fixed. It was a bug in the way function arguments are cast to SQL types. Apparently H2 is a bit more tolerant than PostgreSQL.
Greetings, Sebastian 2014-10-16 10:19 GMT+02:00 Sebastian Schaffert < sebastian.schaff...@gmail.com>: > Hi Fabian, > > I will check, maybe the substring semantics in PostgreSQL is different (it > does have a different syntax). > > Greetings, > > Sebastian > > 2014-10-16 8:33 GMT+02:00 Fabian Cretton <fabian.cret...@hevs.ch>: > >> Sebastian, >> >> I just had a try: >> The query now works on my installation with the default H2 back-end. >> But it seems it doesn't work on my PostrgreSQL installation. >> >> Sorry that I don't have debug information to give you. >> >> From my trial with this query, without the Distinct: >> prefix qb: <http://purl.org/linked-data/cube#> >> prefix sdmx-dimension: <http://purl.org/linked-data/sdmx/2009/dimension#> >> SELECT * WHERE { >> ?obs a qb:Observation; >> sdmx-dimension:timePeriod ?timeStamp. >> BIND (SUBSTR(str(?timeStamp), 1, 4) AS ?year ) >> } limit 10 >> >> the ?year column is empty whereas teh ?timeStamp does contain the correct >> value. >> >> Once again, I am in no hurry about that :-) >> Fabian >> >> >>> Sebastian Schaffert <sebastian.schaff...@gmail.com> 15.10.2014 14:45 >> >>> >> Hi Fabian, >> >> this one was easy: we forgot to implement the substring function. If you >> update from the GIT develop branch it should work now, can you have a try? >> >> Greetings, >> >> Sebastian >> >> 2014-10-15 14:18 GMT+02:00 Sebastian Schaffert < >> sebastian.schaff...@gmail.com>: >> >>> Hi Fabian, >>> >>> DISTINCT is implemented. I need to have a look at the query to see why >>> it is not working in this case. My guess it is because of the BIND (which >>> is awkward to implement in SQL). I am in the middle of preparing for ISWC, >>> so I am not sure I will manage this week if the problem is more complex. :-S >>> >>> A good way to see WHY a SPARQL query is not working is to turn on debug >>> logging for SPARQL, add the debug log appender to SPARQL and then run the >>> query. The marmotta-debug.log file will then contain the SQL produced by >>> the SPARQL query. >>> >>> Greetings, >>> >>> Sebastian >>> >>> 2014-10-15 14:01 GMT+02:00 Fabian Cretton <fabian.cret...@hevs.ch>: >>> >>>> Hi, >>>> I thought DISTINCT was implemented, but it seems not ? >>>> This query gives back all results with duplicates: >>>> prefix qb: <http://purl.org/linked-data/cube#> >>>> prefix sdmx-dimension: <http://purl.org/linked-data/sdmx/2009/dimension >>>> #> >>>> SELECT DISTINCT ?year WHERE { >>>> ?obs a qb:Observation; >>>> sdmx-dimension:timePeriod ?timeStamp. >>>> BIND (SUBSTR(xsd:string(?timeStamp), 1, 4) AS ?year ) >>>> } >>>> (On Sesame-OWLIM it works fine) >>>> thanks for any advice >>>> Fabian >>>> >>> >>> >> >
