Skip to content

[hotfix] Fix union read fail when from timestamp and projection enabled #2383

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

Merged
luoyuxia merged 1 commit into from
Jan 19, 2026
Merged

[hotfix] Fix union read fail when from timestamp and projection enabled #2383

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
4 changes: 2 additions & 2 deletions ...link/fluss-flink-common/src/main/java/org/apache/fluss/flink/lake/LakeSplitGenerator.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -119,7 +119,7 @@ public List<SourceSplitBase> generateHybridLakeFlussSplits() throws Exception {
lakeSplits, isLogTable, tableBucketsOffset, partitionNameById);
} else {
Map<Integer, List<LakeSplit>> nonPartitionLakeSplits =
lakeSplits.values().iterator().next();
lakeSplits.isEmpty() ? null : lakeSplits.values().iterator().next();
Copy link
Contributor Author

@luoyuxia luoyuxia Jan 15, 2026

Choose a reason for hiding this comment

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

when no split is generate, it may be empty. Let's use the safe way.

// non-partitioned table
return generateNoPartitionedTableSplit(
nonPartitionLakeSplits, isLogTable, tableBucketsOffset);
Expand Down Expand Up @@ -307,7 +307,7 @@ private SourceSplitBase generateSplitForPrimaryKeyTableBucket(
}

private List<SourceSplitBase> generateNoPartitionedTableSplit(
Map<Integer, List<LakeSplit>> lakeSplits,
@Nullable Map<Integer, List<LakeSplit>> lakeSplits,
boolean isLogTable,
Map<TableBucket, Long> tableBucketSnapshotLogOffset) {
// iterate all bucket
Expand Down
8 changes: 3 additions & 5 deletions ...link/fluss-flink-common/src/main/java/org/apache/fluss/flink/source/FlinkTableSource.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -331,8 +331,7 @@ public boolean isBounded() {
enableLakeSource = false;
} else {
if (enableLakeSource) {
enableLakeSource =
pushTimeStampFilterToLakeSource(lakeSource, flussRowType);
enableLakeSource = pushTimeStampFilterToLakeSource(lakeSource);
}
}
break;
Expand Down Expand Up @@ -385,12 +384,11 @@ public boolean isBounded() {
}
}

private boolean pushTimeStampFilterToLakeSource(
LakeSource<?> lakeSource, RowType flussRowType) {
private boolean pushTimeStampFilterToLakeSource(LakeSource<?> lakeSource) {
// will push timestamp to lake
// we will have three additional system columns, __bucket, __offset, __timestamp
// in lake, get the __timestamp index in lake table
final int timestampFieldIndex = flussRowType.getFieldCount() + 2;
final int timestampFieldIndex = tableOutputType.getFieldCount() + 2;
Copy link
Contributor Author

@luoyuxia luoyuxia Jan 15, 2026

Choose a reason for hiding this comment

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

flussRowType is projected row, we should use origin tableOutputType

Predicate timestampFilter =
new LeafPredicate(
GreaterOrEqual.INSTANCE,
Expand Down
21 changes: 11 additions & 10 deletions ...-common/src/main/java/org/apache/fluss/flink/source/enumerator/FlinkSourceEnumerator.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -770,16 +770,6 @@ private void assignPendingSplits(Set<Integer> pendingReaders) {
TableBucket tableBucket = split.getTableBucket();
assignedTableBuckets.add(tableBucket);

if (pendingHybridLakeFlussSplits != null) {
// removed from the pendingHybridLakeFlussSplits
// since this split already be assigned
pendingHybridLakeFlussSplits.removeIf(
hybridLakeFlussSplit ->
hybridLakeFlussSplit
.splitId()
.equals(split.splitId()));
}

if (isPartitioned) {
long partitionId =
checkNotNull(
Expand All @@ -792,6 +782,17 @@ private void assignPendingSplits(Set<Integer> pendingReaders) {
assignedPartitions.put(partitionId, partitionName);
}
});

if (pendingHybridLakeFlussSplits != null) {
Copy link
Contributor Author

@luoyuxia luoyuxia Jan 15, 2026

Choose a reason for hiding this comment

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

move out from the loop to speed it.

Set<String> splitIdsToRemove =
pendingAssignmentForReader.stream()
.map(SourceSplitBase::splitId)
.collect(Collectors.toSet());
// removed from the pendingHybridLakeFlussSplits
// since this split already be assigned
pendingHybridLakeFlussSplits.removeIf(
split -> splitIdsToRemove.contains(split.splitId()));
}
}
}

Expand Down
8 changes: 6 additions & 2 deletions ...n/src/test/java/org/apache/fluss/lake/paimon/flink/FlinkUnionReadFromTimestampITCase.java
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -140,12 +140,16 @@ void testUnionReadFromTimestamp() throws Exception {
CloseableIterator<Row> actualRows =
streamTEnv
.executeSql(
"select * from "
"select b from "
+ tableName
+ " /*+ OPTIONS('scan.startup.mode' = 'timestamp',\n"
+ "'scan.startup.timestamp' = '2000') */")
.collect();
List<Row> expectedRows = rows.stream().skip(2 * 3).collect(Collectors.toList());
List<Row> expectedRows =
rows.stream()
.skip(2 * 3)
.map(row -> Row.of(row.getField(1)))
.collect(Collectors.toList());
assertRowResultsIgnoreOrder(actualRows, expectedRows, true);

// verify scan from earliest
Expand Down