You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@marmotta.apache.org by Sebastian Schaffert <se...@gmail.com> on 2016/11/19 13:24:22 UTC

Re: issues loading DBpedia 2015-10 with KiWiLoader

So I got that one. The dataset contains a line

<http://wikidata.dbpedia.org/resource/Q1541> <
http://wikidata.dbpedia.org/property/birthDate> "-0106-01-03"^^<
http://www.w3.org/2001/XMLSchema#date> .

with a negative date that starts with a 0, which is invalid according to
the XMLSchema data datatype (https://www.w3.org/TR/xmlschema11-2/#date)
where a 0 is not allowed as first character of the year. PostgreSQL in any
case doesn't like it.

I am trying to work around the problem.




Sergio Fernández <wi...@apache.org> schrieb am Di., 11. Okt. 2016 um
10:24 Uhr:

> On Tue, Aug 9, 2016 at 9:55 PM, Sebastian Schaffert <
> sebastian.schaffert@gmail.com> wrote:
>
> > Do you have a concrete test file for which it fails? I can try then. Just
> > don't want to download all the DBPedia language versions... ;-)
> >
>
> I think
>
> http://downloads.dbpedia.org/2016-04/core-i18n/ta/infobox_properties_wkd_uris_ta.ttl.bz2
> is the one causing issues loading on KiWi... But I have to double check
> that. It'd be great is someone else could give it a try. ;-)
>
>
> --
> Sergio Fernández
> Partner Technology Manager
> Redlink GmbH
> m: +43 6602747925 <+43%20660%202747925>
> e: sergio.fernandez@redlink.co
> w: http://redlink.co
>

Re: issues loading DBpedia 2015-10 with KiWiLoader

Posted by Sebastian Schaffert <se...@gmail.com>.
Ok, I fixed this, and the dataset now imports.

Note that the problem will be worse in other databases than PostgreSQL.
MySQL doesn't support dates before Christ at all.

Sebastian

Sebastian Schaffert <se...@gmail.com> schrieb am Sa., 19.
Nov. 2016 um 14:24 Uhr:

> So I got that one. The dataset contains a line
>
> <http://wikidata.dbpedia.org/resource/Q1541> <
> http://wikidata.dbpedia.org/property/birthDate> "-0106-01-03"^^<
> http://www.w3.org/2001/XMLSchema#date> .
>
> with a negative date that starts with a 0, which is invalid according to
> the XMLSchema data datatype (https://www.w3.org/TR/xmlschema11-2/#date)
> where a 0 is not allowed as first character of the year. PostgreSQL in any
> case doesn't like it.
>
> I am trying to work around the problem.
>
>
>
>
> Sergio Fernández <wi...@apache.org> schrieb am Di., 11. Okt. 2016 um
> 10:24 Uhr:
>
> On Tue, Aug 9, 2016 at 9:55 PM, Sebastian Schaffert <
> sebastian.schaffert@gmail.com> wrote:
>
> > Do you have a concrete test file for which it fails? I can try then. Just
> > don't want to download all the DBPedia language versions... ;-)
> >
>
> I think
>
> http://downloads.dbpedia.org/2016-04/core-i18n/ta/infobox_properties_wkd_uris_ta.ttl.bz2
> is the one causing issues loading on KiWi... But I have to double check
> that. It'd be great is someone else could give it a try. ;-)
>
>
> --
> Sergio Fernández
> Partner Technology Manager
> Redlink GmbH
> m: +43 6602747925 <+43%20660%202747925>
> e: sergio.fernandez@redlink.co
> w: http://redlink.co
>
>

Re: issues loading DBpedia 2015-10 with KiWiLoader

Posted by Sergio Fernández <se...@redlink.co>.
Hi Sebastian,

On Sat, Nov 19, 2016 at 2:24 PM, Sebastian Schaffert <
sebastian.schaffert@gmail.com> wrote:

> So I got that one. The dataset contains a line
>
> <http://wikidata.dbpedia.org/resource/Q1541> <
> http://wikidata.dbpedia.org/property/birthDate> "-0106-01-03"^^<
> http://www.w3.org/2001/XMLSchema#date> .
>
> with a negative date that starts with a 0, which is invalid according to
> the XMLSchema data datatype (https://www.w3.org/TR/xmlschema11-2/#date)
> where a 0 is not allowed as first character of the year. PostgreSQL in any
> case doesn't like it.
>
> I am trying to work around the problem.
>

I saw the patch:
https://github.com/apache/marmotta/commit/08e6f79d595a3ce22d95e612825bfcbc3f9c6686

Thanks, I'll give it a try.

Honestly I didn't have much time to actually debug the issue. How do you
debug such kind of issues on wrong statements in batches produced by the
loader?

Cheers,


Sergio Fernández <wi...@apache.org> schrieb am Di., 11. Okt. 2016 um
> 10:24 Uhr:
>
> > On Tue, Aug 9, 2016 at 9:55 PM, Sebastian Schaffert <
> > sebastian.schaffert@gmail.com> wrote:
> >
> > > Do you have a concrete test file for which it fails? I can try then.
> Just
> > > don't want to download all the DBPedia language versions... ;-)
> > >
> >
> > I think
> >
> > http://downloads.dbpedia.org/2016-04/core-i18n/ta/infobox_
> properties_wkd_uris_ta.ttl.bz2
> > is the one causing issues loading on KiWi... But I have to double check
> > that. It'd be great is someone else could give it a try. ;-)
> >
> >
> > --
> > Sergio Fernández
> > Partner Technology Manager
> > Redlink GmbH
> > m: +43 6602747925 <+43%20660%202747925>
> > e: sergio.fernandez@redlink.co
> > w: http://redlink.co
> >
>



-- 
Sergio Fernández
Partner Technology Manager
Redlink GmbH
m: +43 6602747925
e: sergio.fernandez@redlink.co
w: http://redlink.co