Uploaded image for project: 'Hive'
  1. Hive
  2. HIVE-6551

group by after join with skew join optimization references invalid task sometimes

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Resolved
    • Trivial
    • Resolution: Fixed
    • None
    • 0.13.0
    • None
    • None

    Description

      For example,

      hive> set hive.auto.convert.join = true;
      hive> set hive.optimize.skewjoin = true;
      hive> set hive.skewjoin.key = 3;
      hive> 
          > EXPLAIN FROM 
          > (SELECT src.* FROM src) x
          > JOIN 
          > (SELECT src.* FROM src) Y
          > ON (x.key = Y.key)
          > SELECT sum(hash(Y.key)), sum(hash(Y.value));
      OK
      STAGE DEPENDENCIES:
        Stage-8 is a root stage
        Stage-6 depends on stages: Stage-8
        Stage-5 depends on stages: Stage-6 , consists of Stage-4, Stage-2
        Stage-4
        Stage-2 depends on stages: Stage-4, Stage-1
        Stage-0 is a root stage
      ...
      

      Stage-2 references not-existing Stage-1

      Attachments

        1. HIVE-6551.1.patch.txt
          3 kB
          Navis Ryu

        Activity

          People

            navis Navis Ryu
            navis Navis Ryu
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: