[jira] [Created] (JENA-1257) Bad performance with VALUES blocks

["Osma Suominen (JIRA)" <ji...@apache.org>]

Osma Suominen created JENA-1257:
-----------------------------------

             Summary: Bad performance with VALUES blocks
                 Key: JENA-1257
                 URL: https://issues.apache.org/jira/browse/JENA-1257
             Project: Apache Jena
          Issue Type: Bug
          Components: ARQ
    Affects Versions: Jena 3.1.1
         Environment: Ubuntu 12.04 LTS amd64
java version "1.8.0_91" (Oracle)

            Reporter: Osma Suominen
         Attachments: explain.zip

I'm having performance issues with a query involving VALUES blocks. It seems to me that the values are bound to variables too late, which causes long execution times.

Also discussed on users list: http://mail-archives.apache.org/mod_mbox/jena-users/201611.mbox/%3c58185A49.7060100@helsinki.fi%3e

Steps to repeat:

1. Create a TDB with YSO:

{noformat}
wget http://api.finto.fi/download/yso/yso-skos.ttl
tdbloader --loc YSO yso-skos.ttl
{noformat}

2. Run the below query against the TDB:

{noformat}
tdbquery --loc YSO --query ysoslowunion.rq
{noformat}

Contents of ysoslowunion.rq:
{noformat}
PREFIX skos: <http://www.w3.org/2004/02/skos/core#>
SELECT *
WHERE {
  VALUES ?uri { <http://www.yso.fi/onto/yso/p864> }
  { ?s ?p ?uri }
  UNION
  { ?uri ?p ?o 
    OPTIONAL {
      ?x skos:member ?o .
      FILTER NOT EXISTS {
        ?x skos:member ?other .
        FILTER NOT EXISTS {
          ?other skos:broader ?uri
        }
      }
    }
  }
} 
{noformat}

With Jena 3.1.0 this takes <1 second.
With current 3.1.1 snapshot it takes 12 seconds.

Possibly related to changes made in JENA-1167.

I will attach --explain output for both versions as a zip. The 3.1.1 output is 473 MB uncompressed.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)