You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@geode.apache.org by "Jason Huynh (JIRA)" <ji...@apache.org> on 2017/03/07 19:29:38 UTC

[jira] [Created] (GEODE-2615) Limit should not be applied at the index level for certain queries

Jason Huynh created GEODE-2615:
----------------------------------

             Summary: Limit should not be applied at the index level for certain queries
                 Key: GEODE-2615
                 URL: https://issues.apache.org/jira/browse/GEODE-2615
             Project: Geode
          Issue Type: Bug
          Components: querying
            Reporter: Jason Huynh


There are certain queries that are having limit applied at the index level that are then returning incorrect/missing results.

An example of one type of query:
elect distinct * from /regionA regionA,/regionB regionB where regionA.code = regionB.code and regionB.someField in Set('0','1','2') and regionA.otherField = '8' and (regionA.code='3' or regionA.name like '%') limit 1

This query ends up applying the limit from the index with 1 result but when joined with the other region, filters out the 1 result and ends up with 0 results.  We should probably be able to detect if this query uses multiple indexes or if there are other operands and decided not to limit at the index level, or remove limit at the index level overall (similar to how range index no longer applies limit at the index level)



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)