Issues

• Remove the spill replay workaround from the memory pools once DataFusion includes apache/datafusion#25383

  area:aggregationarea:memoryenhancementrequires-triage
  apache/datafusion-comet#6583 · Filter by author andygrove opened on Oct 3, 2026
  

• Native Iceberg Write: close cleanup ownership gap during JVM handoff

  enhancementrequires-triage
  apache/datafusion-comet#6581 · Filter by author unikdahal opened on Oct 3, 2026
  1

  

• My priorities for 1.2.0

  enhancementrequires-triage
  apache/datafusion-comet#6580 · Filter by author andygrove opened on Oct 3, 2026 · 1.2.0
  

• Arrow struct writer should respect projected schema width

  bugrequires-triage
  apache/datafusion-comet#6576 · Filter by author unikdahal opened on Oct 3, 2026
  1

  

• Native scans resolve the AWS SDK ProfileCredentialsProvider names differently from Hadoop

  requires-triage
  apache/datafusion-comet#6575 · Filter by author dwsmith1983 opened on Oct 3, 2026

• input_file_name() returns empty values above a converted Spark Parquet scan

  bugrequires-triage
  apache/datafusion-comet#6573 · Filter by author andygrove opened on Oct 3, 2026

• EXPLAIN omits the cached plan under CometInMemoryTableScan

  bugrequires-triage
  apache/datafusion-comet#6572 · Filter by author comphead opened on Oct 3, 2026
  

• Codegen dispatcher initializes kernels with an incorrect partition index under UNION ALL and coalesce

  bugrequires-triage
  apache/datafusion-comet#6570 · Filter by author LinSimon-901101 opened on Oct 3, 2026

• [EPIC] Faster Spark-to-Arrow conversion in CometSparkToColumnarExec

  area:scanenhancementEPICperformancerequires-triage
  apache/datafusion-comet#6565 · Filter by author andygrove opened on Oct 3, 2026
  

• Native Iceberg writer counts NaNs under NULL structs and NULL list or map entries

  area:Icebergarea:writerbugpriority:mediumrequires-triage
  apache/datafusion-comet#6562 · Filter by author andygrove opened on Oct 3, 2026

• Native rebase refusals raise CometNativeException instead of SparkUpgradeException

  requires-triage
  apache/datafusion-comet#6558 · Filter by author dwsmith1983 opened on Oct 3, 2026

• Support OneRowRelation as a native Comet source

  requires-triage
  apache/datafusion-comet#6555 · Filter by author pingzh opened on Oct 2, 2026
  1

  

• Comet 1.2.0 Release

  enhancementrequires-triage
  apache/datafusion-comet#6550 · Filter by author andygrove opened on Oct 2, 2026 · 1.2.0
  

• Native map construction doesn't match Spark 4.0+ float key normalization (-0.0 keys, missing DUPLICATED_MAP_KEY)

  bugrequires-triage
  apache/datafusion-comet#6549 · Filter by author andygrove opened on Oct 2, 2026

• S3 credential SPI: follow-ups from #6478 (hostless Iceberg metadata locations, docs, tests)

  area:Icebergbugdocumentationrequires-triagetest
  apache/datafusion-comet#6536 · Filter by author andygrove opened on Oct 2, 2026

• Native CASE WHEN and COALESCE reconcile struct fields by name instead of position

  bugcorrectnessrequires-triage
  apache/datafusion-comet#6532 · Filter by author andygrove opened on Oct 1, 2026

• AQE skew join split makes the join fall back to Spark with no fallback reason

  performancerequires-triage
  apache/datafusion-comet#6530 · Filter by author comphead opened on Oct 1, 2026
  1

  

• perf: Comet hash joins use more task time than Spark (skewed SHJ with nested payload, BHJ)

  performancerequires-triage
  apache/datafusion-comet#6528 · Filter by author comphead opened on Oct 1, 2026
  

• Reduce retained buffer allocation when map lookups select short nested values

  area:expressionsbugpriority:mediumrequires-triage
  apache/datafusion-comet#6525 · Filter by author andygrove opened on Oct 1, 2026
  1

  

• Native Iceberg sorted-merge scan rebuilds the delete-file cache per file (shared deletes re-read)

  bug
  apache/datafusion-comet#6524 · Filter by author parthchandra opened on Oct 1, 2026
  

• Clarify how to run native Comet -> Celeborn shuffle end-to-end

  requires-triage
  apache/datafusion-comet#6523 · Filter by author unikdahal opened on Oct 1, 2026
  6

• Run array_contains on float elements natively with Spark's equality instead of through the codegen dispatcher

  enhancementrequires-triage
  apache/datafusion-comet#6520 · Filter by author andygrove opened on Oct 1, 2026
  1

  

• percentile_approx returns different percentiles from Spark when its input holds a NaN with the sign bit set

  bugrequires-triage
  apache/datafusion-comet#6519 · Filter by author andygrove opened on Oct 1, 2026
  1

  

• Support native existence sort-merge joins

  requires-triage
  apache/datafusion-comet#6514 · Filter by author coderfender opened on Oct 1, 2026
  

• Native Parquet scan assigns row groups to splits differently from Spark

  area:scanenhancementrequires-triage
  apache/datafusion-comet#6512 · Filter by author andygrove opened on Oct 1, 2026 · 1.2.0
  1