First off I recognize this is probably more of an H2 issue than ignite but I'm asking her in hopes someone has seen similar behavior
I have Partitioned cache on which I'm trying to run the following query. Keys are collocated based on the internalRobotName My target query is: "SELECT S1.* FROM STATEFULROBOTTELEMETRY S1 where S1.internalRobotName = ? and internalMissionId = (SELECT internalMissionId FROM STATEFULROBOTTELEMETRY S2 WHERE S2.internalRobotName = ? ORDER BY internalTimeStamp DESC LIMIT 1 ) I have grouped index like (internalRobotName , internalTimeStamp, internalMissionId ) The above query with a moderate amount of keys ~1M runs very slowly. Essentially I'm trying to find the latest grouping of entries for what my teams calls a mission. I believe that the subquery should use the index to push the limit down and return only the latest missionId. If instead I first query for the latest mission, essentially pulling the subquery out into a intermediate cache query, then pass that as an arg via query.setArgs(...) to a final query, then that runs quite bit faster and meets performance criteria. Any thoughts on this? -barrett -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Correlated-Subquery-Performance-tp5877.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
