Skip to content

[AURON #2434] Support full-data-file Iceberg changelog deletes #2435

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
weimingdiit wants to merge 3 commits into
base: master
Choose a base branch
from
+100 −36
Open

[AURON #2434] Support full-data-file Iceberg changelog deletes #2435

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
5f8f024
[AURON #2434] Support full-data-file Iceberg changelog deletes
weimingdiit Jul 27, 2026
c0fea01
update code
weimingdiit Jul 30, 2026
f3db3df
fix checkstyle
weimingdiit Jul 30, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
60 changes: 38 additions & 22 deletions .../auron-iceberg/src/main/scala/org/apache/spark/sql/auron/iceberg/IcebergScanSupport.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,7 +22,7 @@ import scala.collection.JavaConverters._
import scala.util.control.NonFatal

import org.apache.commons.lang3.reflect.MethodUtils
import org.apache.iceberg.{AddedRowsScanTask, ChangelogOperation, ChangelogScanTask, FileFormat, FileScanTask, MetadataColumns, ScanTask}
import org.apache.iceberg.{AddedRowsScanTask, ChangelogOperation, ChangelogScanTask, DataFile, DeletedDataFileScanTask, FileFormat, FileScanTask, MetadataColumns, ScanTask}
import org.apache.iceberg.expressions.{And => IcebergAnd, BoundPredicate, Expression => IcebergExpression, Not => IcebergNot, Or => IcebergOr, UnboundPredicate}
import org.apache.iceberg.spark.source.AuronIcebergSourceUtil
import org.apache.spark.internal.Logging
Expand Down Expand Up @@ -267,22 +267,14 @@ object IcebergScanSupport extends Logging {
return None
}

val addedRowsTasks = changelogTasks.collect { case task: AddedRowsScanTask => task }
// First native changelog support is insert-only. Delete and update images need Iceberg
// delete-file handling, so keep them on Spark's reader for now.
if (addedRowsTasks.size != changelogTasks.size) {
// Native changelog scan can read data-file rows directly for insert tasks and
// full-data-file delete tasks. Row-level deletes still need Iceberg delete-file handling.
val nativeChangelogTasks = changelogTasks.flatMap(toNativeChangelogDataFileTask)
if (nativeChangelogTasks.size != changelogTasks.size) {
return None
}

if (!addedRowsTasks.forall(_.operation() == ChangelogOperation.INSERT)) {
return None
}

if (!addedRowsTasks.forall(task => deletesEmpty(task.deletes()))) {
return None
}

val formats = addedRowsTasks.map(_.file().format()).distinct
val formats = nativeChangelogTasks.map(_.file.format()).distinct
if (formats.size > 1) {
return None
}
Expand All @@ -298,7 +290,7 @@ object IcebergScanSupport extends Logging {
}

val pruningPredicates = collectPruningPredicates(scan.asInstanceOf[AnyRef], readSchema)
val nativeTasks = addedRowsTasks.map(task => toNativeScanTask(task, partitionSchema))
val nativeTasks = nativeChangelogTasks.map(task => toNativeScanTask(task, partitionSchema))
Some(
IcebergScanPlan(
nativeTasks,
Expand Down Expand Up @@ -528,6 +520,12 @@ object IcebergScanSupport extends Logging {

private case class IcebergPartitionView(tasks: Seq[ScanTask])

private case class NativeChangelogDataFileTask(
file: DataFile,
start: Long,
length: Long,
changelogTask: ChangelogScanTask)

private def icebergPartition(partition: InputPartition): Option[IcebergPartitionView] = {
val className = partition.getClass.getName
// Only accept Iceberg SparkInputPartition to access task groups.
Expand Down Expand Up @@ -559,6 +557,21 @@ object IcebergScanSupport extends Logging {
}
}

private def toNativeChangelogDataFileTask(
task: ChangelogScanTask): Option[NativeChangelogDataFileTask] = {
task match {
case added: AddedRowsScanTask
if added.operation() == ChangelogOperation.INSERT &&
deletesEmpty(added.deletes()) =>
Some(NativeChangelogDataFileTask(added.file(), added.start(), added.length(), added))
case deleted: DeletedDataFileScanTask if deletesEmpty(deleted.existingDeletes()) =>
Some(
NativeChangelogDataFileTask(deleted.file(), deleted.start(), deleted.length(), deleted))
case _ =>
None
}
}
Comment on lines +560 to +573

@weiqingy weiqingy Aug 1, 2026

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No change requested, just evidence on this one.

The changelog task type that would actually be unsafe to run natively is DeletedRowsScanTask, the merge-on-read sibling that carries addedDeletes(). In Iceberg 1.10.1 it declares (javadoc stripped):

public interface DeletedRowsScanTask extends ChangelogScanTask, ContentScanTask<DataFile> {
  List<DeleteFile> addedDeletes();

  List<DeleteFile> existingDeletes();

  @Override
  default ChangelogOperation operation() {
    return ChangelogOperation.DELETE;
  }
}

It extends ChangelogScanTask, not DeletedDataFileScanTask, so the type match on this line is what keeps it off the native path. Its operation() is hard-coded DELETE as well, so an operation() check would not have excluded it either. Only the type match does.

For the _change_type angle: the metadata case at IcebergScanSupport.scala:620 reads the value straight off the task (requiredChangelogTask(name).operation().name()) rather than off the Scala type, so a subtype that overrode operation() would stamp its own value instead of a mismatched one.

@weimingdiit weimingdiit Aug 4, 2026

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, that matches my reading.

The native path is intentionally limited by the concrete task type here. DeletedRowsScanTask is a separate ChangelogScanTask implementation, not a DeletedDataFileScanTask, so it will not match this case and will continue to fall back. Keeping or removing the operation() == DELETE check does not affect that, since DeletedRowsScanTask also hard-codes operation() to DELETE.

For _change_type, I agree the value should come from the task operation itself. The metadata materialization path already uses requiredChangelogTask(name).operation().name(), so if Iceberg ever provides a supported changelog task with a different operation value, the metadata column will reflect the task value rather than the Scala match type.

No additional code change from my side.


private def toNativeScanTask(
task: FileScanTask,
partitionSchema: StructType): IcebergNativeScanTask = {
Expand All @@ -572,15 +585,18 @@ object IcebergScanSupport extends Logging {
}

private def toNativeScanTask(
task: AddedRowsScanTask,
task: NativeChangelogDataFileTask,
partitionSchema: StructType): IcebergNativeScanTask = {
val file = task.file()
IcebergNativeScanTask(
file.location(),
task.start(),
task.length(),
file.fileSizeInBytes(),
metadataPartitionValues(file.location(), file.specId(), Some(task), partitionSchema))
task.file.location(),
task.start,
task.length,
task.file.fileSizeInBytes(),
metadataPartitionValues(
task.file.location(),
task.file.specId(),
Some(task.changelogTask),
partitionSchema))
}

private def metadataPartitionValues(
Expand Down
76 changes: 62 additions & 14 deletions .../auron-iceberg/src/test/scala/org/apache/auron/iceberg/AuronIcebergIntegrationSuite.scala
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -597,6 +597,48 @@ class AuronIcebergIntegrationSuite
}
}

test("iceberg native scan supports full-data-file delete changelog scan") {
withTable("local.db.t_changelog_full_file_delete") {
withTempView("t_changelog_full_file_delete_changes") {
sql("""
|create table local.db.t_changelog_full_file_delete (id int, v string, p int)
|using iceberg
|partitioned by (p)
|tblproperties ('format-version' = '2')
|""".stripMargin)
sql("""
|insert into local.db.t_changelog_full_file_delete
|values (1, 'a', 1), (2, 'b', 2)
|""".stripMargin)
val startSnapshotId = currentSnapshotId("local.db.t_changelog_full_file_delete")
sql("delete from local.db.t_changelog_full_file_delete where p = 1")
val endSnapshotId = currentSnapshotId("local.db.t_changelog_full_file_delete")
createChangelogView(
"local.db.t_changelog_full_file_delete",
"t_changelog_full_file_delete_changes",
startSnapshotId,
endSnapshotId)

val query =
"""
|select id, v, p, _change_type
|from t_changelog_full_file_delete_changes
|order by id
|""".stripMargin
val expected = Seq(Row(1, "a", 1, "DELETE"))
withSQLConf("spark.auron.enable" -> "false") {
checkAnswer(sql(query), expected)
}
withSQLConf("spark.auron.enable" -> "true", "spark.auron.enable.iceberg.scan" -> "true") {
val df = sql(query)
checkAnswer(df, expected)
val nativeScan = executedNativeIcebergTableScanExec(df)
assert(nativeScan.staticPlan.scanTasks.size == 1)
}
}
}
}

test("iceberg native changelog scan remains correct in dynamic pruning join") {
withTable("local.db.t_changelog_dpp", "local.db.t_changelog_dpp_dim") {
withTempView("t_changelog_dpp_changes") {
Expand Down Expand Up @@ -700,39 +742,45 @@ class AuronIcebergIntegrationSuite
}
}

test("iceberg changelog scan falls back when delete changes exist") {
withTable("local.db.t_changelog_delete") {
withTempView("t_changelog_delete_changes") {
test("iceberg native scan supports mixed insert and full-data-file delete changelog scan") {
Comment thread
weimingdiit marked this conversation as resolved.
withTable("local.db.t_changelog_mixed_delete") {
withTempView("t_changelog_mixed_delete_changes") {
sql("""
|create table local.db.t_changelog_delete (id int, v string)
|create table local.db.t_changelog_mixed_delete (id int, v string, p int)
Comment on lines +745 to +749
|using iceberg
|partitioned by (p)
|tblproperties ('format-version' = '2')
|""".stripMargin)
sql("insert into local.db.t_changelog_delete values (1, 'a'), (2, 'b')")
val startSnapshotId = currentSnapshotId("local.db.t_changelog_delete")
sql("delete from local.db.t_changelog_delete where id = 1")
val endSnapshotId = currentSnapshotId("local.db.t_changelog_delete")
sql("""
|insert into local.db.t_changelog_mixed_delete
|values (1, 'a', 1), (2, 'b', 2)
|""".stripMargin)
val startSnapshotId = currentSnapshotId("local.db.t_changelog_mixed_delete")
sql("delete from local.db.t_changelog_mixed_delete where p = 1")
sql("insert into local.db.t_changelog_mixed_delete values (3, 'c', 3)")
val endSnapshotId = currentSnapshotId("local.db.t_changelog_mixed_delete")
createChangelogView(
"local.db.t_changelog_delete",
"t_changelog_delete_changes",
"local.db.t_changelog_mixed_delete",
"t_changelog_mixed_delete_changes",
startSnapshotId,
endSnapshotId)

val query =
"""
|select id, v, _change_type, _change_ordinal, _commit_snapshot_id
|from t_changelog_delete_changes
|select id, v, p, _change_type, _change_ordinal, _commit_snapshot_id
|from t_changelog_mixed_delete_changes
|order by id, _change_type
|""".stripMargin
var expected: Seq[Row] = Nil
withSQLConf("spark.auron.enable" -> "false") {
expected = sql(query).collect().toSeq
}
assert(expected.exists(row => row.getString(3) == "DELETE"))
assert(expected.exists(row => row.getString(3) == "INSERT"))
withSQLConf("spark.auron.enable" -> "true", "spark.auron.enable.iceberg.scan" -> "true") {
val df = sql(query)
checkAnswer(df, expected)
val plan = df.queryExecution.executedPlan.toString()
assert(!plan.contains("NativeIcebergTableScan"))
executedNativeIcebergTableScanExec(df)
}
}
}
Expand Down
Loading