[jira] [Created] (RYA-374) Aggregation Result Join Bug

["Caleb Meier (JIRA)" <ji...@apache.org>]

Caleb Meier created RYA-374:
-------------------------------

             Summary: Aggregation Result Join Bug
                 Key: RYA-374
                 URL: https://issues.apache.org/jira/browse/RYA-374
             Project: Rya
          Issue Type: Bug
          Components: clients
    Affects Versions: 3.2.11
            Reporter: Caleb Meier
            Assignee: Caleb Meier


Right now, it is not possible to join on aggregation results that come from nested queries within the Rya Trigger Service.  For example, in the following query

select ?msg ?lastObserved where {
   {select (MAX(?time) AS ?lastObserved) {
        ?msg <uri:hasTime> ?time
   }}
   ?msg <uri:hasTime> ?lastObserved
}

would not return the msg with the last observed time.  It returns all messages, paired with
the last returned time.  So it does not join on the last observed time.  This is because of how aggregation results are handled in the trigger service.  The values are stored in the value field of the key, as opposed to the row.  When a join is performed, the Fluo table is scanned using a prefix over the contents of its rows.  Therefore, the aggregation result is not taken into account.  



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)