Inaccurate num keys scanned estimate for trivial queries

XMLWordPrintableJSON

    • Type: Bug
    • Resolution: Gone away
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: None
    • Query Optimization
    • ALL
    • Hide 
      db.adminCommand({setParameter:1, featureFlagCostBasedRanker: true, internalQueryCBRCEMode: "samplingCE"});
db.a.drop()
db.a.createIndex({a: 1})
db.a.insertOne({a: [1, 2]})
db.a.explain().find({}).sort({a: 1})
      Show
      db.adminCommand({setParameter:1, featureFlagCostBasedRanker: true , internalQueryCBRCEMode: "samplingCE" }); db.a.drop() db.a.createIndex({a: 1}) db.a.insertOne({a: [1, 2]}) db.a.explain().find({}).sort({a: 1})
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      With SERVER-108882, we switch to heuristicce for trivial queries.

      For trivially true match expressions planned over a multikey index, the resulting number of keys scanned estimate is inaccurate.

      This does not cause any problems for access path selection within the scope of CBR, but a future project using CBR may run into this limitation.

            Assignee:
            Unassigned
            Reporter:
            Kartal Kaan Bozdogan
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: