Uploaded image for project: 'Spark'
  1. Spark
  2. SPARK-32542

Add an optimizer rule to split an Expand into multiple Expands for aggregates

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Resolved
    • Major
    • Resolution: Won't Fix
    • 3.0.0
    • None
    • SQL
    • None

    Description

      Split an expand into several small Expand, which contains the Specified number of projections.
      For instance, like this sql.select a, b, c, d, count(1) from table1 group by a, b, c, d with cube. It can expand 2^4 times of original data size.
      Now we specify the spark.sql.optimizer.projections.size=4.The Expand will be split into 2^4/4 smallExpand.It can reduce the shuffle pressure and improve performance in multidimensional analysis when data is huge.

      Attachments

        Activity

          People

            Unassigned Unassigned
            wangshikai karl wang
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: