Uploaded image for project: 'IMPALA'
  1. IMPALA
  2. IMPALA-6621

Improve In-predicate performance by using an alternative data structure for checking set membership

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Resolved
    • Major
    • Resolution: Fixed
    • Impala 2.13.0
    • Impala 2.13.0
    • Backend
    • None
    • ghx-label-4

    Description

      Currently when using a SET_LOOKUP strategy for in-predicates in impala we use an std:set object for checking membership.
      Using other data structures like boost::unordered_set and boost::flat_set we can get a significant performance improvement. Please see attached results of micro benchmarks using std::set, flat_set and unordered_set.

      Attachments

        1. release_build_BoostFlatset.txt.txt
          57 kB
          Bikramjeet Vig
        2. release_build_BoostUnorderedset.txt.txt
          57 kB
          Bikramjeet Vig
        3. release_build_StdSet.txt.txt
          57 kB
          Bikramjeet Vig

        Issue Links

          Activity

            People

              bikramjeet.vig Bikramjeet Vig
              bikramjeet.vig Bikramjeet Vig
              Votes:
              0 Vote for this issue
              Watchers:
              6 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: