Ok, so we may want to add a little xdmp:sleep() after the document insert
instruction I guess. Thanks for your help.

cheers,
Jakob.


On Tue, May 6, 2014 at 5:53 PM, John Snelson <[email protected]>wrote:

> Triple index updates are transactional, so you can see them as soon as
> the inserting transaction has completed. I suspect you're seeing a race
> conditions with the inserting transaction.
>
> John
>
> On 06/05/14 16:22, Jakob Fix wrote:
> > John,
> >
> > Maybe a misunderstanding in my message: We're inserting the XML document
> > which triggers the pre-commit trigger that generates the triple document
> > in a *separate transaction from the test itself*, in order to be able to
> > query it. For this to work, we're inserting the document via an HTTP
> > POST request of our API, a call that is blocking, so we expect the
> > document to be query-able (which is the case), and a SPARQL query to
> > return the inserted triples (which is not always the case).
> >
> >
> > cheers,
> > Jakob.
> >
> >
> > On Tue, May 6, 2014 at 4:21 PM, John Snelson <[email protected]
> > <mailto:[email protected]>> wrote:
> >
> >     On 06/05/14 15:17, Jakob Fix wrote:
> >      > we are doing xray unit testing for our application, and one of the
> >      > issues we run into occasionally is that even though we know that
> >     triples
> >      > have been created and stored as triple documents, they are not
> always
> >      > (but sometimes) found when executing a sparql query in the same
> >     test. by
> >      > the way, triples are created by extracting information from an XML
> >      > document, via a pre-commit trigger before storing the XML document
> >      > proper. oh yes, and the document is inserted in a separate
> >     transaction.
> >      >
> >      > could this be due to the fact that these documents are not indexed
> >      > immediately, but periodically?
> >
> >     No, it's because you insert the document containing triples in a
> >     separate transaction. Why don't you insert it in the same
> transaction if
> >     you're using pre-commit triggers?
> >
> >     John
> >
> >     --
> >     John Snelson, Lead Engineer http://twitter.com/jpcs
> >     MarkLogic Corporation http://www.marklogic.com
> >     _______________________________________________
> >     General mailing list
> >     [email protected] <mailto:
> [email protected]>
> >     http://developer.marklogic.com/mailman/listinfo/general
> >
> >
> >
> >
> > _______________________________________________
> > General mailing list
> > [email protected]
> > http://developer.marklogic.com/mailman/listinfo/general
> >
>
>
> --
> John Snelson, Lead Engineer                    http://twitter.com/jpcs
> MarkLogic Corporation                         http://www.marklogic.com
> _______________________________________________
> General mailing list
> [email protected]
> http://developer.marklogic.com/mailman/listinfo/general
>
_______________________________________________
General mailing list
[email protected]
http://developer.marklogic.com/mailman/listinfo/general

Reply via email to