Pull requests · apache/datafusion-comet

feat: [DO NOT MERGE] native Iceberg reads of unshredded Variant columns

#7575mbutrovich opened

Nov 9, 2026

Contributor

refactor: factor out map-entry hashing loops into `hash_map_entries!`

#7569jinjin1919 opened

Nov 8, 2026

feat: warn about unrecognized spark.comet configuration keys

#7568tlfua opened

Nov 8, 2026

fix: warn about the deprecated memoryPool.fraction when it is set on the session

#7567tlfua opened

Nov 8, 2026

fix: share the grouping key type checks between doConvert and canAggregateBeConverted

#7566tlfua opened

Nov 8, 2026

fix: keep URL passwords and query strings out of rejected path reasons

#7565dwsmith1983 opened

Nov 7, 2026

Contributor

fix: keep URL query strings and passwords out of native Delta errors

#7562dwsmith1983 opened

Nov 7, 2026

Contributor

fix: keep URL passwords and query strings out of Parquet writer output

#7561dwsmith1983 opened

Nov 7, 2026

Contributor

chore: remove the Spark 3.4 shim source sets

#7560andygrove opened

Nov 7, 2026

MemberQueued

1

fix: keep URL passwords and query strings out of object store errors

#7556dwsmith1983 opened

Nov 7, 2026

Contributor

fix: sort a Spark window's doubted input by the whole order it reports

#7555dwsmith1983 opened

Nov 7, 2026

Contributor

fix: reuse the Comet broadcast for a non-AQE DPP subquery whose build side is not a CometNativeExec

#7554tlfua opened

Nov 7, 2026

fix: cast a non-INT `GetArrayItem` index to INT like Spark

#7552akupchinskiy opened

Nov 7, 2026

Contributor

2

feat: run hll_sketch_estimate natively by default and narrow HLL incompatibility reasons

#7550tlfua opened

Nov 7, 2026

fix: support integer seed for rand and randn

#7547galaxytemple opened

Nov 7, 2026

fix: bound and validate requests to the Comet Native Executors cluster

#7546andygrove opened

Nov 6, 2026

Member

docs: add research notes on hybrid execution for Comet Native Executors

#7545andygrove opened

Nov 6, 2026

Member

feat: add the Comet Native Executors driver (experimental)

#7544andygrove opened

Nov 6, 2026

Member

10

perf: let a partial aggregate skip aggregation for AVG

#7543andygrove opened

Nov 6, 2026

Member

ci: let committers bypass the merge queue for urgent fixes

#7540comphead opened

Nov 6, 2026

Contributor

1

docs: allow justified slower shapes in expression optimizations, add code guidelines

#7539mbutrovich opened

Nov 6, 2026

Contributor

fix: split shuffle blocks whose dictionary columns would expand past 2 GiB

#7537ted-jenks opened

Nov 6, 2026

fix: have the native Iceberg write read a converted child's Arrow stream

#7536zhangfengcdt opened

Nov 6, 2026

Member

refactor: deduplicate spark_ceil and spark_floor

#7534FreakingMind opened

Nov 6, 2026

2

build(deps): bump parquet from 59.3.1 to 59.3.2 in /native/native-executors

#7533dependabot[bot] opened

Nov 6, 2026

Bot

Navigation Menu

Repository navigation

All pull requests

feat: [DO NOT MERGE] native Iceberg reads of unshredded Variant columns

refactor: factor out map-entry hashing loops into `hash_map_entries!`

feat: warn about unrecognized spark.comet configuration keys

fix: warn about the deprecated memoryPool.fraction when it is set on the session

fix: share the grouping key type checks between doConvert and canAggregateBeConverted

fix: keep URL passwords and query strings out of rejected path reasons

fix: keep URL query strings and passwords out of native Delta errors

fix: keep URL passwords and query strings out of Parquet writer output

chore: remove the Spark 3.4 shim source sets

fix: keep URL passwords and query strings out of object store errors

fix: sort a Spark window's doubted input by the whole order it reports

fix: reuse the Comet broadcast for a non-AQE DPP subquery whose build side is not a CometNativeExec

fix: cast a non-INT `GetArrayItem` index to INT like Spark

feat: run hll_sketch_estimate natively by default and narrow HLL incompatibility reasons

fix: support integer seed for rand and randn

fix: bound and validate requests to the Comet Native Executors cluster

docs: add research notes on hybrid execution for Comet Native Executors

feat: add the Comet Native Executors driver (experimental)

perf: let a partial aggregate skip aggregation for AVG

ci: let committers bypass the merge queue for urgent fixes

docs: allow justified slower shapes in expression optimizations, add code guidelines

fix: split shuffle blocks whose dictionary columns would expand past 2 GiB

fix: have the native Iceberg write read a converted child's Arrow stream

refactor: deduplicate spark_ceil and spark_floor

build(deps): bump parquet from 59.3.1 to 59.3.2 in /native/native-executors