feat: Per-Partition Plan Building for Native Iceberg Scans [iceberg-rust]#3295

Closed

andygrove wants to merge 23 commits into

apache:mainfrom

andygrove:iceberg-split-serialization

Commits on Jan 27, 2026

[EXPERIMENTAL] Split Iceberg FileScanTask serialization for reduced network transfer

andygrove
and
claude
committed
Change iceberg-rust CI jobs to trigger on [iceberg-rust] tag

andygrove
and
claude
committed
Allow clippy::large_enum_variant for spark_operator module

andygrove
and
claude
committed
trigger CI
andygrove
committed
Remove unused import

andygrove
and
claude
committed
Simplify split mode detection for Iceberg scan serialization

andygrove
and
claude
committed
Revert "Simplify split mode detection for Iceberg scan serialization"
andygrove
committed
Fix partition index handling in IcebergScanExec for split mode

andygrove
and
claude
committed
Keep legacy code path for IcebergScan as child of other operators

andygrove
and
claude
committed
Implement per-partition plan injection for Iceberg split serialization

andygrove
and
claude
committed
Remove unused legacy apply method from CometIcebergNativeScanExec

andygrove
and
claude
committed
Remove split_mode flag and simplify IcebergScan protobuf

andygrove
and
claude
committed
Remove PR_DESCRIPTION.md

andygrove
and
claude
committed
fix: Use perPartitionData.length for Iceberg partition count

andygrove
and
claude
committed
fix: Override convertBlock() to preserve transient fields

andygrove
and
claude
committed
style: Remove unused KeyGroupedPartitioning import
andygrove
committed
trigger CI
andygrove
committed
Handle partition index out of bounds for joins/multi-input scenarios

andygrove
and
claude
committed
Revert "Handle partition index out of bounds for joins/multi-input scenarios"
andygrove
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

feat: Per-Partition Plan Building for Native Iceberg Scans [iceberg-rust]#3295

feat: Per-Partition Plan Building for Native Iceberg Scans [iceberg-rust]#3295
andygrove wants to merge 23 commits into
apache:mainfrom
andygrove:iceberg-split-serialization

Commits on Jan 27, 2026

[EXPERIMENTAL] Split Iceberg FileScanTask serialization for reduced network transfer

Change iceberg-rust CI jobs to trigger on [iceberg-rust] tag

Allow clippy::large_enum_variant for spark_operator module

trigger CI

Remove unused import

Simplify split mode detection for Iceberg scan serialization

Revert "Simplify split mode detection for Iceberg scan serialization"

Fix partition index handling in IcebergScanExec for split mode

Keep legacy code path for IcebergScan as child of other operators

Implement per-partition plan injection for Iceberg split serialization

Remove unused legacy apply method from CometIcebergNativeScanExec

Remove split_mode flag and simplify IcebergScan protobuf

Remove PR_DESCRIPTION.md

fix: Use perPartitionData.length for Iceberg partition count

fix: Override convertBlock() to preserve transient fields

style: Remove unused KeyGroupedPartitioning import

trigger CI

Handle partition index out of bounds for joins/multi-input scenarios

Revert "Handle partition index out of bounds for joins/multi-input scenarios"

Commits on Jan 28, 2026

fix: Handle joins over multiple Iceberg tables in split serialization

scalastyle

chore: Remove debug logging from CometIcebergSplitRDD

fix: Handle broadcast joins over Iceberg tables in split serialization