[GitHub] [jena] rvesse commented on issue #1634: Sorting fails with custom datatype literal

[GitBox <gi...@apache.org>]

rvesse commented on issue #1634:
URL: https://github.com/apache/jena/issues/1634#issuecomment-1326519765

   > By the way, what's the difference in sorting with and without projecting the sorted variable? With selection of `?l` it doesn't fail
   
   So the SPARQL specification [section 18.2.5](https://www.w3.org/TR/sparql11-query/#convertSolMod) says that `ORDER BY` is applied prior to `DISTINCT`
   
   **BUT** ARQ includes an optimiser that swaps the ordering i.e. applies `DISTINCT` first in cases where all the variables in the `ORDER BY` are also projected.  This is semantically equivalent and is shown to substantially increase performance because you throw out all the non-distinct solutions prior to ordering.  So in that case every row is guaranteed to be unique and you can't ever get any unstable sorting results as every row is different (at least for your dataset).
   
   I suspect what happens in the failing case is that you have two solutions with the custom datatype literals present and we aren't guaranteeing to provide a stable sort (see PR #1406 that Andy referenced) over those.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: issues-unsubscribe@jena.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@jena.apache.org
For additional commands, e-mail: issues-help@jena.apache.org